typo

some programs are now 100% source compatible between C64 and Cx16 targets!
import libraries have been rena;med

README.md

@@ -50,7 +50,7 @@ What use Prog8 provide?
 
 - "c64": Commodore-64  (6510 CPU = almost a 6502)  premium support.
 - "cx16": [CommanderX16](https://www.commanderx16.com)  (65c02 CPU)  experimental support.
-
+- If you only use standard kernel and prog8 library routines, it is possible to compile the *exact same program* for both machines (just change the compiler target flag)!
 
 
 
@@ -75,16 +75,18 @@ Example code
 
 This code calculates prime numbers using the Sieve of Eratosthenes algorithm::
 
-    %import c64textio
+    %import textio
     %zeropage basicsafe
     
     main {
     
         ubyte[256] sieve
-        ubyte candidate_prime = 2
+        ubyte candidate_prime = 2       ; is increased in the loop
     
         sub start() {
-            memset(sieve, 256, false)   ; clear the sieve
+            ; clear the sieve, to reset starting situation on subsequent runs
+            memset(sieve, 256, false)   
+            ; calculate primes
             txt.print("prime numbers up to 255:\n\n")
             ubyte amount=0
             repeat {
@@ -95,18 +97,19 @@ This code calculates prime numbers using the Sieve of Eratosthenes algorithm::
                 txt.print(", ")
                 amount++
             }
-            c64.CHROUT('\n')
+            txt.chrout('\n')
             txt.print("number of primes (expected 54): ")
             txt.print_ub(amount)
-            c64.CHROUT('\n')
+            txt.chrout('\n')
         }
     
         sub find_next_prime() -> ubyte {
             while sieve[candidate_prime] {
                 candidate_prime++
                 if candidate_prime==0
-                    return 0        ; we wrapped; no more primes available
+                    return 0        ; we wrapped; no more primes available in the sieve
             }
+    
             ; found next one, mark the multiples and return it.
             sieve[candidate_prime] = true
             uword multiple = candidate_prime
@@ -120,11 +123,11 @@ This code calculates prime numbers using the Sieve of Eratosthenes algorithm::
     }
 
 
+
 when compiled an ran on a C-64 you'll get:
 
 ![c64 screen](docs/source/_static/primes_example.png)
 
-
 One of the included examples (wizzine.p8) animates a bunch of sprite balloons and looks like this:
 
 ![wizzine screen](docs/source/_static/wizzine.png)
@@ -136,3 +139,8 @@ Another example (cube3d-sprites.p8) draws the vertices of a rotating 3d cube:
 If you want to play a video game, a fully working Tetris clone is included in the examples:
 
 ![tehtriz_screen](docs/source/_static/tehtriz.png)
+
+The CommanderX16 compiler target is quite capable already too, here's a well known space ship 
+animated in 3D with hidden line removal, in the CommanderX16 emulator:
+
+![cobra3d](docs/source/_static/cobra3d.png)

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/c64/floats.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
+		lda  #<FL_ONE_const
-		ldy  #>FL_FONE
+		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
+		stx  P8ZP_SCRATCH_REG
-		lda  #<FL_FONE
+		lda  #<FL_ONE_const
-		ldy  #>FL_FONE
+		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
+		lda  #<FL_ONE_const
-		ldy  #>FL_FONE
+		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
+		lda  #<FL_ZERO_const
-		ldy  #>ZERO
+		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/c64/floats.p8

@@ -4,14 +4,14 @@
 ;
 ; indent format: TABS, size=8
 
+%target c64
 %option enable_floats
 
-c64flt {
+floats {
 	; ---- this block contains C-64 floating point related functions ----
 
         const float  PI     = 3.141592653589793
         const float  TWOPI  = 6.283185307179586
-        const  float  ZERO      = 0.0
 
 
 ; ---- C64 basic and kernal ROM float constants and functions ----
@@ -50,22 +50,22 @@ romsub $bc0f = MOVEF() clobbers(A,X)                        ; copy fac1 to fac2
 romsub $bbd4 = MOVMF(uword mflpt @ XY) clobbers(A,Y)        ; store fac1 to memory  X/Y as 5-byte mflpt
 
 ; fac1-> signed word in Y/A (might throw ILLEGAL QUANTITY)
-; (tip: use c64flt.FTOSWRDAY to get A/Y output; lo/hi switched to normal little endian order)
+; (tip: use floats.FTOSWRDAY to get A/Y output; lo/hi switched to normal little endian order)
 romsub $b1aa = FTOSWORDYA() clobbers(X) -> ubyte @ Y, ubyte @ A       ; note: calls AYINT.
 
 ; fac1 -> unsigned word in Y/A (might throw ILLEGAL QUANTITY) (result also in $14/15)
-; (tip: use c64flt.GETADRAY to get A/Y output; lo/hi switched to normal little endian order)
+; (tip: use floats.GETADRAY to get A/Y output; lo/hi switched to normal little endian order)
 romsub $b7f7 = GETADR() clobbers(X) -> ubyte @ Y, ubyte @ A
 
 romsub $bc9b = QINT() clobbers(A,X,Y)           ; fac1 -> 4-byte signed integer in 98-101 ($62-$65), with the MSB FIRST.
 romsub $b1bf = AYINT() clobbers(A,X,Y)          ; fac1-> signed word in 100-101 ($64-$65) MSB FIRST. (might throw ILLEGAL QUANTITY)
 
 ; GIVAYF: signed word in Y/A (note different lsb/msb order) -> float in fac1
-; (tip: use c64flt.GIVAYFAY to use A/Y input; lo/hi switched to normal order)
+; (tip: use floats.GIVAYFAY to use A/Y input; lo/hi switched to normal order)
-; there is also c64flt.GIVUAYFAY - unsigned word in A/Y (lo/hi) to fac1
+; there is also floats.GIVUAYFAY - unsigned word in A/Y (lo/hi) to fac1
-; there is also c64flt.FREADS32  that reads from 98-101 ($62-$65) MSB FIRST
+; there is also floats.FREADS32  that reads from 98-101 ($62-$65) MSB FIRST
-; there is also c64flt.FREADUS32  that reads from 98-101 ($62-$65) MSB FIRST
+; there is also floats.FREADUS32  that reads from 98-101 ($62-$65) MSB FIRST
-; there is also c64flt.FREADS24AXY  that reads signed int24 into fac1 from A/X/Y (lo/mid/hi bytes)
+; there is also floats.FREADS24AXY  that reads signed int24 into fac1 from A/X/Y (lo/mid/hi bytes)
 romsub $b391 = GIVAYF(ubyte lo @ Y, ubyte hi @ A) clobbers(A,X,Y)
 
 romsub $b3a2 = FREADUY(ubyte value @ Y) clobbers(A,X,Y)     ; 8 bit unsigned Y -> float in fac1
@@ -194,24 +194,24 @@ 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
+		sta  P8ZP_SCRATCH_W1
-		sty  P8ZP_SCRATCH_REG
+		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
 	}}
 }
 
-%asminclude "library:c64floats.asm", ""
+%asminclude "library:c64/floats.asm", ""
 
 }

compiler/res/prog8lib/c64/graphics.p8

@@ -1,11 +1,14 @@
-%import c64textio
+%target c64
+%import textio
 
 ; bitmap pixel graphics module for the C64
-; only black/white monchrome for now
+; only black/white monchrome 320x200 for now
 ; assumes bitmap screen memory is $2000-$3fff
 
 graphics {
-    const uword bitmap_address = $2000
+    const uword BITMAP_ADDRESS = $2000
+    const uword WIDTH = 320
+    const ubyte HEIGHT = 200
 
     sub enable_bitmap_mode() {
         ; enable bitmap screen, erase it and set colors to black/white.
@@ -15,7 +18,7 @@ graphics {
     }
 
     sub clear_screen(ubyte pixelcolor, ubyte bgcolor) {
-        memset(bitmap_address, 320*200/8, 0)
+        memset(BITMAP_ADDRESS, 320*200/8, 0)
         txt.fill_screen(pixelcolor << 4 | bgcolor, 0)
     }
 
@@ -134,33 +137,33 @@ graphics {
         }
     }
 
-    sub disc(uword cx, ubyte cy, ubyte radius) {
+    sub disc(uword xcenter, ubyte ycenter, ubyte radius) {
         ; Midpoint algorithm, filled
         ubyte xx = radius
         ubyte yy = 0
         byte decisionOver2 = 1-xx as byte
 
         while xx>=yy {
-            ubyte cy_plus_yy = cy + yy
+            ubyte ycenter_plus_yy = ycenter + yy
-            ubyte cy_min_yy = cy - yy
+            ubyte ycenter_min_yy = ycenter - yy
-            ubyte cy_plus_xx = cy + xx
+            ubyte ycenter_plus_xx = ycenter + xx
-            ubyte cy_min_xx = cy - xx
+            ubyte ycenter_min_xx = ycenter - xx
 
-            for internal_plotx in cx to cx+xx {
+            for internal_plotx in xcenter to xcenter+xx {
-                internal_plot(cy_plus_yy)
+                internal_plot(ycenter_plus_yy)
-                internal_plot(cy_min_yy)
+                internal_plot(ycenter_min_yy)
             }
-            for internal_plotx in cx-xx to cx-1 {
+            for internal_plotx in xcenter-xx to xcenter-1 {
-                internal_plot(cy_plus_yy)
+                internal_plot(ycenter_plus_yy)
-                internal_plot(cy_min_yy)
+                internal_plot(ycenter_min_yy)
             }
-            for internal_plotx in cx to cx+yy {
+            for internal_plotx in xcenter to xcenter+yy {
-                internal_plot(cy_plus_xx)
+                internal_plot(ycenter_plus_xx)
-                internal_plot(cy_min_xx)
+                internal_plot(ycenter_min_xx)
             }
-            for internal_plotx in cx-yy to cx {
+            for internal_plotx in xcenter-yy to xcenter {
-                internal_plot(cy_plus_xx)
+                internal_plot(ycenter_plus_xx)
-                internal_plot(cy_min_xx)
+                internal_plot(ycenter_min_xx)
             }
             yy++
             if decisionOver2<=0
@@ -176,13 +179,11 @@ graphics {
 ; here is the non-asm code for the plot routine below:
 ;    sub plot_nonasm(uword px, ubyte py) {
 ;        ubyte[] ormask = [128, 64, 32, 16, 8, 4, 2, 1]
-;        uword addr = bitmap_address + 320*(py>>3) + (py & 7) + (px & %0000000111111000)
+;        uword addr = BITMAP_ADDRESS + 320*(py>>3) + (py & 7) + (px & %0000000111111000)
 ;        @(addr) |= ormask[lsb(px) & 7]
 ;    }
 
-    ; TODO fix the use of X (or XY) as parameter so we can actually use this plot() routine
+    asmsub  plot(uword plotx @XY, ubyte ploty @A) clobbers (A, X, Y) {
-    ;   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) {
         %asm {{
             stx  internal_plotx
             sty  internal_plotx+1
@@ -190,12 +191,14 @@ 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
@@ -219,15 +222,13 @@ 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
 
 ; note: this can be even faster if we also have a 256 byte x-lookup table, but hey.
 ; see http://codebase64.org/doku.php?id=base:various_techniques_to_calculate_adresses_fast_common_screen_formats_for_pixel_graphics
-; the y lookup tables encodes this formula:  bitmap_address + 320*(py>>3) + (py & 7)    (y from 0..199)
+; the y lookup tables encodes this formula:  BITMAP_ADDRESS + 320*(py>>3) + (py & 7)    (y from 0..199)
 ; We use the 64tass syntax for range expressions to calculate this table on assembly time.
 
 _plot_y_values := $2000 + 320*(range(200)>>3) + (range(200) & 7)

compiler/res/prog8lib/c64/syslib.p8

@@ -5,11 +5,13 @@
 ;
 ; indent format: TABS, size=8
 
+%target c64
 
 c64 {
         &ubyte  TIME_HI         = $a0       ; software jiffy clock, hi byte
         &ubyte  TIME_MID        = $a1       ;  .. mid byte
         &ubyte  TIME_LO         = $a2       ;    .. lo byte. Updated by IRQ every 1/60 sec
+        &ubyte  STATUS          = $90       ; kernel status variable for I/O
         &ubyte  STKEY           = $91       ; various keyboard statuses (updated by IRQ)
         &ubyte  SFDX            = $cb       ; current key pressed (matrix value) (updated by IRQ)
 
@@ -258,6 +260,16 @@ asmsub  init_system()  {
     }}
 }
 
+asmsub  reset_system()  {
+    ; Soft-reset the system back to Basic prompt.
+    %asm {{
+        sei
+        lda  #14
+        sta  $01        ; bank the kernal in
+        jmp  (c64.RESET_VEC)
+    }}
+}
+
 asmsub  set_irqvec_excl() clobbers(A)  {
 	%asm {{
 		sei
@@ -298,8 +310,6 @@ _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 +334,6 @@ _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 +348,6 @@ _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/c64/textio.p8

@@ -4,22 +4,22 @@
 ;
 ; indent format: TABS, size=8
 
-
+%target c64
-%import c64lib
+%import syslib
 %import conv
 
 
 txt {
 
-asmsub  clear_screen() {
+const ubyte DEFAULT_WIDTH = 40
-    %asm {{
+const ubyte DEFAULT_HEIGHT = 25
-        lda  #' '
+
-        jmp  clear_screenchars
+
-    }}
+sub  clear_screen() {
+    clear_screenchars(' ')
 }
 
-
+asmsub  fill_screen (ubyte char @ A, ubyte color @ Y) clobbers(A)  {
-asmsub  fill_screen (ubyte char @ A, ubyte charcolor @ Y) clobbers(A)  {
 	; ---- fill the character screen with the given fill character and character color.
 	;      (assumes screen and color matrix are at their default addresses)
 
@@ -49,7 +49,7 @@ _loop		sta  c64.Screen,y
         }}
 }
 
-asmsub  clear_screencolors (ubyte scrcolor @ A) clobbers(Y)  {
+asmsub  clear_screencolors (ubyte color @ A) clobbers(Y)  {
 	; ---- clear the character screen colors with the given color (leaves characters).
 	;      (assumes color matrix is at the default address)
 	%asm {{
@@ -68,13 +68,21 @@ sub color (ubyte txtcol) {
     c64.COLOR = txtcol
 }
 
-asmsub  scroll_left_full  (ubyte alsocolors @ Pc) clobbers(A, Y)  {
+sub lowercase() {
+    c64.VMCSB |= 2
+}
+
+sub uppercase() {
+    c64.VMCSB &= ~2
+}
+
+asmsub  scroll_left  (ubyte alsocolors @ Pc) clobbers(A, Y)  {
 	; ---- scroll the whole screen 1 character to the left
 	;      contents of the rightmost 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
 
@@ -102,17 +110,17 @@ _scroll_screen  ; scroll the screen memory
 		dey
 		bpl  -
 
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
 
-asmsub  scroll_right_full  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_right  (ubyte alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character to the right
 	;      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
 
@@ -136,17 +144,17 @@ _scroll_screen  ; scroll the screen memory
 		dex
 		bpl  -
 
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
 
-asmsub  scroll_up_full  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_up  (ubyte alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character up
 	;      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
 
@@ -170,17 +178,17 @@ _scroll_screen  ; scroll the screen memory
 		dex
 		bpl  -
 
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
 
-asmsub  scroll_down_full  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_down  (ubyte alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character down
 	;      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
 
@@ -204,11 +212,13 @@ _scroll_screen  ; scroll the screen memory
 		dex
 		bpl  -
 
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
 
+romsub $FFD2 = chrout(ubyte char @ A)    ; for consistency. You can also use c64.CHROUT directly ofcourse.
+
 asmsub  print (str text @ AY) clobbers(A,Y)  {
 	; ---- print null terminated string from A/Y
 	; note: the compiler contains an optimization that will replace
@@ -230,7 +240,7 @@ 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 +249,7 @@ 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 +257,7 @@ 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 +274,7 @@ _print_byte_digits
         jsr  c64.CHROUT
 _ones   txa
 		jsr  c64.CHROUT
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
@@ -272,7 +282,7 @@ _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 +290,14 @@ asmsub  print_b  (byte value @ A) clobbers(A,Y)  {
 		jsr  c64.CHROUT
 +		pla
 		jsr  conv.byte2decimal
-		jsr  print_ub._print_byte_digits
+		jmp  print_ub._print_byte_digits
-		ldx  P8ZP_SCRATCH_REG_X
-		rts
 	}}
 }
 
 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 +307,7 @@ 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 +315,7 @@ 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 +328,7 @@ 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 +361,7 @@ 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 +369,7 @@ 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 +377,9 @@ 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
@@ -434,21 +442,22 @@ asmsub  input_chars  (uword buffer @ AY) clobbers(A) -> ubyte @ Y  {
 	}}
 }
 
-asmsub  setchr  (ubyte col @Y, ubyte row @A) clobbers(A)  {
+asmsub  setchr  (ubyte col @X, ubyte row @Y, ubyte character @A) clobbers(A, Y)  {
-	; ---- set the character in SCRATCH_ZPB1 on the screen matrix at the given position
+	; ---- sets the character in the screen matrix at the given position
 	%asm {{
-		sty  P8ZP_SCRATCH_REG
+		pha
+		tya
 		asl  a
 		tay
 		lda  _screenrows+1,y
 		sta  _mod+2
-		lda  _screenrows,y
+		txa
 		clc
-		adc  P8ZP_SCRATCH_REG
+		adc  _screenrows,y
 		sta  _mod+1
 		bcc  +
 		inc  _mod+2
-+		lda  P8ZP_SCRATCH_B1
++		pla
 _mod		sta  $ffff		; modified
 		rts
 
@@ -456,17 +465,18 @@ _screenrows	.word  $0400 + range(0, 1000, 40)
 	}}
 }
 
-asmsub  getchr  (ubyte col @Y, ubyte row @A) clobbers(Y) -> ubyte @ A {
+asmsub  getchr  (ubyte col @A, ubyte row @Y) clobbers(Y) -> ubyte @ A {
 	; ---- get the character in the screen matrix at the given location
 	%asm  {{
-		sty  P8ZP_SCRATCH_B1
+		pha
+		tya
 		asl  a
 		tay
 		lda  setchr._screenrows+1,y
 		sta  _mod+2
-		lda  setchr._screenrows,y
+		pla
 		clc
-		adc  P8ZP_SCRATCH_B1
+		adc  setchr._screenrows,y
 		sta  _mod+1
 		bcc  _mod
 		inc  _mod+2
@@ -475,21 +485,22 @@ _mod		lda  $ffff		; modified
 	}}
 }
 
-asmsub  setclr  (ubyte col @Y, ubyte row @A) clobbers(A)  {
+asmsub  setclr  (ubyte col @X, ubyte row @Y, ubyte color @A) clobbers(A, Y)  {
-	; ---- set the color in SCRATCH_ZPB1 on the screen matrix at the given position
+	; ---- set the color in A on the screen matrix at the given position
 	%asm {{
-		sty  P8ZP_SCRATCH_REG
+		pha
+		tya
 		asl  a
 		tay
 		lda  _colorrows+1,y
 		sta  _mod+2
-		lda  _colorrows,y
+		txa
 		clc
-		adc  P8ZP_SCRATCH_REG
+		adc  _colorrows,y
 		sta  _mod+1
 		bcc  +
 		inc  _mod+2
-+		lda  P8ZP_SCRATCH_B1
++		pla
 _mod		sta  $ffff		; modified
 		rts
 
@@ -497,17 +508,18 @@ _colorrows	.word  $d800 + range(0, 1000, 40)
 	}}
 }
 
-asmsub  getclr  (ubyte col @Y, ubyte row @A) clobbers(Y) -> ubyte @ A {
+asmsub  getclr  (ubyte col @A, ubyte row @Y) clobbers(Y) -> ubyte @ A {
 	; ---- get the color in the screen color matrix at the given location
 	%asm  {{
-		sty  P8ZP_SCRATCH_B1
+		pha
+		tya
 		asl  a
 		tay
 		lda  setclr._colorrows+1,y
 		sta  _mod+2
-		lda  setclr._colorrows,y
+		pla
 		clc
-		adc  P8ZP_SCRATCH_B1
+		adc  setclr._colorrows,y
 		sta  _mod+1
 		bcc  _mod
 		inc  _mod+2
@@ -545,11 +557,29 @@ _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
+	}}
+}
+
+asmsub width() clobbers(X,Y) -> ubyte @A {
+    ; -- returns the text screen width (number of columns)
+    %asm {{
+        jsr  c64.SCREEN
+        txa
+        rts
+    }}
+}
+
+asmsub height() clobbers(X, Y) -> ubyte @A {
+    ; -- returns the text screen height (number of rows)
+    %asm {{
+        jsr  c64.SCREEN
+        tya
         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/cx16/floats.p8

@@ -4,14 +4,14 @@
 ;
 ; indent format: TABS, size=8
 
+%target cx16
 %option enable_floats
 
-c64flt {
+floats {
 	; ---- this block contains C-64 floating point related functions ----
 
         const float  PI     = 3.141592653589793
         const float  TWOPI  = 6.283185307179586
-        const  float  ZERO      = 0.0
 
 
 ; ---- ROM float functions ----
@@ -25,7 +25,7 @@ c64flt {
 romsub $fe00 = AYINT() clobbers(A,X,Y)          ; fac1-> signed word in 100-101 ($64-$65) MSB FIRST. (might throw ILLEGAL QUANTITY)
 
 ; GIVAYF: signed word in Y/A (note different lsb/msb order) -> float in fac1
-; there is also c64flt.GIVUAYFAY - unsigned word in A/Y (lo/hi) to fac1
+; there is also floats.GIVUAYFAY - unsigned word in A/Y (lo/hi) to fac1
 ; (tip: use GIVAYFAY to use A/Y input; lo/hi switched to normal order)
 romsub $fe03 = GIVAYF(ubyte lo @ Y, ubyte hi @ A) clobbers(A,X,Y)
 
@@ -130,24 +130,24 @@ 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
+		sta  P8ZP_SCRATCH_W1
-		sty  P8ZP_SCRATCH_REG
+		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
+		rts
 	}}
 }
 
-%asminclude "library:c64floats.asm", ""
+%asminclude "library:c64/floats.asm", ""
 
 }

compiler/res/prog8lib/cx16/graphics.p8

@@ -0,0 +1,156 @@
+%target cx16
+%import syslib
+
+; bitmap pixel graphics module for the CommanderX16
+; wraps the graphics functions that are in ROM.
+; only black/white monchrome 320x200 for now.
+
+graphics {
+    const uword WIDTH = 320
+    const ubyte HEIGHT = 200
+
+    sub enable_bitmap_mode() {
+        ; enable bitmap screen, erase it and set colors to black/white.
+        void cx16.screen_set_mode($80)
+        cx16.r0 = 0
+        cx16.GRAPH_init()
+        clear_screen(1, 0)
+    }
+
+    sub clear_screen(ubyte pixelcolor, ubyte bgcolor) {
+        cx16.GRAPH_set_colors(pixelcolor, pixelcolor, bgcolor)
+        cx16.GRAPH_clear()
+    }
+
+    sub line(uword @zp x1, ubyte @zp y1, uword @zp x2, ubyte @zp y2) {
+        cx16.r0 = x1
+        cx16.r1 = y1
+        cx16.r2 = x2
+        cx16.r3 = y2
+        cx16.GRAPH_draw_line()
+    }
+
+    sub circle(uword xcenter, ubyte ycenter, ubyte radius) {
+        ;cx16.r0 = xcenter - radius/2
+        ;cx16.r1 = ycenter - radius/2
+        ;cx16.r2 = radius*2
+        ;cx16.r3 = radius*2
+        ;cx16.GRAPH_draw_oval(false)          ; TODO  currently is not implemented on cx16, does a BRK
+
+        ; Midpoint algorithm
+        ubyte @zp xx = radius
+        ubyte @zp yy = 0
+        byte @zp decisionOver2 = 1-xx as byte
+
+        while xx>=yy {
+            cx16.r0 = xcenter + xx
+            cx16.r1 = ycenter + yy
+            cx16.FB_cursor_position()
+            cx16.FB_set_pixel(1)
+            cx16.r0 = xcenter - xx
+            cx16.FB_cursor_position()
+            cx16.FB_set_pixel(1)
+            cx16.r0 = xcenter + xx
+            cx16.r1 = ycenter - yy
+            cx16.FB_cursor_position()
+            cx16.FB_set_pixel(1)
+            cx16.r0 = xcenter - xx
+            cx16.FB_cursor_position()
+            cx16.FB_set_pixel(1)
+            cx16.r0 = xcenter + yy
+            cx16.r1 = ycenter + xx
+            cx16.FB_cursor_position()
+            cx16.FB_set_pixel(1)
+            cx16.r0 = xcenter - yy
+            cx16.FB_cursor_position()
+            cx16.FB_set_pixel(1)
+            cx16.r0 = xcenter + yy
+            cx16.r1 = ycenter - xx
+            cx16.FB_cursor_position()
+            cx16.FB_set_pixel(1)
+            cx16.r0 = xcenter - yy
+            cx16.FB_cursor_position()
+            cx16.FB_set_pixel(1)
+            yy++
+            if decisionOver2<=0 {
+                decisionOver2 += 2*yy+1
+            } else {
+                xx--
+                decisionOver2 += 2*(yy-xx)+1
+            }
+        }
+    }
+
+    sub disc(uword xcenter, ubyte ycenter, ubyte radius) {
+;        cx16.r0 = xcenter - radius/2
+;        cx16.r1 = ycenter - radius/2
+;        cx16.r2 = radius*2
+;        cx16.r3 = radius*2
+;        cx16.GRAPH_draw_oval(true)          ; TODO  currently is not implemented on cx16, does a BRK
+
+        ubyte xx = radius
+        ubyte yy = 0
+        byte decisionOver2 = 1-xx as byte
+
+        while xx>=yy {
+            ubyte ycenter_plus_yy = ycenter + yy
+            ubyte ycenter_min_yy = ycenter - yy
+            ubyte ycenter_plus_xx = ycenter + xx
+            ubyte ycenter_min_xx = ycenter - xx
+            uword @zp plotx
+
+            for plotx in xcenter to xcenter+xx {
+                cx16.r0 = plotx
+                cx16.r1 = ycenter_plus_yy
+                cx16.FB_cursor_position()
+                cx16.FB_set_pixel(1)
+                cx16.r1 = ycenter_min_yy
+                cx16.FB_cursor_position()
+                cx16.FB_set_pixel(1)
+            }
+            for plotx in xcenter-xx to xcenter-1 {
+                cx16.r0 = plotx
+                cx16.r1 = ycenter_plus_yy
+                cx16.FB_cursor_position()
+                cx16.FB_set_pixel(1)
+                cx16.r1 = ycenter_min_yy
+                cx16.FB_cursor_position()
+                cx16.FB_set_pixel(1)
+            }
+            for plotx in xcenter to xcenter+yy {
+                cx16.r0 = plotx
+                cx16.r1 = ycenter_plus_xx
+                cx16.FB_cursor_position()
+                cx16.FB_set_pixel(1)
+                cx16.r1 = ycenter_min_xx
+                cx16.FB_cursor_position()
+                cx16.FB_set_pixel(1)
+            }
+            for plotx in xcenter-yy to xcenter {
+                cx16.r0 = plotx
+                cx16.r1 = ycenter_plus_xx
+                cx16.FB_cursor_position()
+                cx16.FB_set_pixel(1)
+                cx16.r1 = ycenter_min_xx
+                cx16.FB_cursor_position()
+                cx16.FB_set_pixel(1)
+            }
+            yy++
+            if decisionOver2<=0
+                decisionOver2 += 2*yy+1
+            else {
+                xx--
+                decisionOver2 += 2*(yy-xx)+1
+            }
+        }
+    }
+
+    sub  plot(uword plotx, ubyte ploty) {
+        cx16.r0 = plotx
+        cx16.r1 = ploty
+        cx16.FB_cursor_position()
+        cx16.FB_set_pixel(1)
+    }
+}
+
+

compiler/res/prog8lib/cx16/syslib.p8

@@ -0,0 +1,291 @@
+; Prog8 definitions for the CommanderX16
+; Including memory registers, I/O registers, Basic and Kernal subroutines.
+;
+; Written by Irmen de Jong (irmen@razorvine.net) - license: GNU GPL 3.0
+;
+; indent format: TABS, size=8
+
+%target cx16
+
+
+c64 {
+
+; ---- kernal routines, these are the same as on the Commodore-64 (hence the same block name) ----
+
+; STROUT --> use screen.print
+; 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 $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 $FF99 = MEMTOP(uword address @ XY, ubyte dir @ Pc) -> ubyte @A, uword @ XY     ; read/set top of memory  pointer, returns number of banks in A
+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 $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 $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 $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 $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
+
+}
+
+cx16 {
+
+; 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  = $0002
+    &uword r1  = $0004
+    &uword r2  = $0006
+    &uword r3  = $0008
+    &uword r4  = $000a
+    &uword r5  = $000c
+    &uword r6  = $000e
+    &uword r7  = $0010
+    &uword r8  = $0012
+    &uword r9  = $0014
+    &uword r10 = $0016
+    &uword r11 = $0018
+    &uword r12 = $001a
+    &uword r13 = $001c
+    &uword r14 = $001e
+    &uword r15 = $0020
+
+; VERA registers
+
+    const uword VERA_BASE = $9F20
+    &ubyte  VERA_ADDR_L   	  = VERA_BASE + $0000
+    &ubyte  VERA_ADDR_M   	  = VERA_BASE + $0001
+    &ubyte  VERA_ADDR_H   	  = VERA_BASE + $0002
+    &ubyte  VERA_DATA0        = VERA_BASE + $0003
+    &ubyte  VERA_DATA1        = VERA_BASE + $0004
+    &ubyte  VERA_CTRL         = VERA_BASE + $0005
+    &ubyte  VERA_IEN          = VERA_BASE + $0006
+    &ubyte  VERA_ISR          = VERA_BASE + $0007
+    &ubyte  VERA_IRQ_LINE_L   = VERA_BASE + $0008
+    &ubyte  VERA_DC_VIDEO     = VERA_BASE + $0009
+    &ubyte  VERA_DC_HSCALE    = VERA_BASE + $000A
+    &ubyte  VERA_DC_VSCALE    = VERA_BASE + $000B
+    &ubyte  VERA_DC_BORDER    = VERA_BASE + $000C
+    &ubyte  VERA_DC_HSTART    = VERA_BASE + $0009
+    &ubyte  VERA_DC_HSTOP     = VERA_BASE + $000A
+    &ubyte  VERA_DC_VSTART    = VERA_BASE + $000B
+    &ubyte  VERA_DC_VSTOP     = VERA_BASE + $000C
+    &ubyte  VERA_L0_CONFIG    = VERA_BASE + $000D
+    &ubyte  VERA_L0_MAPBASE   = VERA_BASE + $000E
+    &ubyte  VERA_L0_TILEBASE  = VERA_BASE + $000F
+    &ubyte  VERA_L0_HSCROLL_L = VERA_BASE + $0010
+    &ubyte  VERA_L0_HSCROLL_H = VERA_BASE + $0011
+    &ubyte  VERA_L0_VSCROLL_L = VERA_BASE + $0012
+    &ubyte  VERA_L0_VSCROLL_H = VERA_BASE + $0013
+    &ubyte  VERA_L1_CONFIG    = VERA_BASE + $0014
+    &ubyte  VERA_L1_MAPBASE   = VERA_BASE + $0015
+    &ubyte  VERA_L1_TILEBASE  = VERA_BASE + $0016
+    &ubyte  VERA_L1_HSCROLL_L = VERA_BASE + $0017
+    &ubyte  VERA_L1_HSCROLL_H = VERA_BASE + $0018
+    &ubyte  VERA_L1_VSCROLL_L = VERA_BASE + $0019
+    &ubyte  VERA_L1_VSCROLL_H = VERA_BASE + $001A
+    &ubyte  VERA_AUDIO_CTRL   = VERA_BASE + $001B
+    &ubyte  VERA_AUDIO_RATE   = VERA_BASE + $001C
+    &ubyte  VERA_AUDIO_DATA   = VERA_BASE + $001D
+    &ubyte  VERA_SPI_DATA     = VERA_BASE + $001E
+    &ubyte  VERA_SPI_CTRL     = VERA_BASE + $001F
+; 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()  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()  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
+
+; high level graphics & fonts
+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()  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)  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()  clobbers(A,X,Y)
+
+
+
+; ---- end of kernal routines ----
+
+asmsub init_system()  {
+    ; Initializes the machine to a sane starting state.
+    ; Called automatically by the loader program logic.
+    %asm {{
+        sei
+        cld
+        ;stz  $00
+        ;stz  $01
+        ;stz  d1prb      ; select rom bank 0
+        lda  #$80
+        sta  VERA_CTRL
+        jsr  c64.IOINIT
+        jsr  c64.RESTOR
+        jsr  c64.CINT
+        lda  #$90       ; black
+        jsr  c64.CHROUT
+        lda  #1         ; swap fg/bg
+        jsr  c64.CHROUT
+        lda  #$9e       ; yellow
+        jsr  c64.CHROUT
+        lda  #147       ; clear screen
+        jsr  c64.CHROUT
+        lda  #0
+        tax
+        tay
+        clc
+        clv
+        cli
+        rts
+    }}
+}
+
+asmsub  reset_system()  {
+    ; Soft-reset the system back to Basic prompt.
+    %asm {{
+        sei
+        lda  #14
+        sta  $01
+        stz  cx16.d1prb         ; bank the kernal in
+        jmp  (cx16.RESET_VEC)
+    }}
+}
+
+}

compiler/res/prog8lib/cx16/textio.p8

@@ -0,0 +1,696 @@
+; Prog8 definitions for the Text I/O and Screen routines for the CommanderX16
+;
+; Written by Irmen de Jong (irmen@razorvine.net) - license: GNU GPL 3.0
+;
+; indent format: TABS, size=8
+
+%target cx16
+%import syslib
+%import conv
+
+
+txt {
+
+const ubyte DEFAULT_WIDTH = 80
+const ubyte DEFAULT_HEIGHT = 60
+
+
+sub  clear_screen() {
+    clear_screenchars(' ')
+}
+
+asmsub  fill_screen (ubyte char @ A, ubyte color @ Y) clobbers(A)  {
+	; ---- fill the character screen with the given fill character and character color.
+	%asm {{
+	    sty  _ly+1
+        phx
+        pha
+        jsr  c64.SCREEN             ; get dimensions in X/Y
+        txa
+        lsr  a
+        lsr  a
+        sta  _lx+1
+        stz  cx16.VERA_CTRL
+        lda  #%00010000
+        sta  cx16.VERA_ADDR_H       ; enable auto increment by 1, bank 0.
+        stz  cx16.VERA_ADDR_L       ; start at (0,0)
+        stz  cx16.VERA_ADDR_M
+        pla
+_lx     ldx  #0                     ; modified
+        phy
+_ly     ldy  #1                     ; modified
+-       sta  cx16.VERA_DATA0
+        sty  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA0
+        sty  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA0
+        sty  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA0
+        sty  cx16.VERA_DATA0
+        dex
+        bne  -
+        ply
+        dey
+        beq  +
+        stz  cx16.VERA_ADDR_L
+        inc  cx16.VERA_ADDR_M       ; next line
+        bra  _lx
++       plx
+        rts
+        }}
+}
+
+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 {{
+        phx
+        pha
+        jsr  c64.SCREEN             ; get dimensions in X/Y
+        txa
+        lsr  a
+        lsr  a
+        sta  _lx+1
+        stz  cx16.VERA_CTRL
+        lda  #%00100000
+        sta  cx16.VERA_ADDR_H       ; enable auto increment by 2, bank 0.
+        stz  cx16.VERA_ADDR_L       ; start at (0,0)
+        stz  cx16.VERA_ADDR_M
+        pla
+_lx     ldx  #0                     ; modified
+-       sta  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA0
+        dex
+        bne  -
+        dey
+        beq  +
+        stz  cx16.VERA_ADDR_L
+        inc  cx16.VERA_ADDR_M       ; next line
+        bra  _lx
++       plx
+        rts
+        }}
+}
+
+asmsub  clear_screencolors (ubyte color @ A) clobbers(Y)  {
+	; ---- clear the character screen colors with the given color (leaves characters).
+	;      (assumes color matrix is at the default address)
+	%asm {{
+        phx
+        sta  _la+1
+        jsr  c64.SCREEN             ; get dimensions in X/Y
+        txa
+        lsr  a
+        lsr  a
+        sta  _lx+1
+        stz  cx16.VERA_CTRL
+        lda  #%00100000
+        sta  cx16.VERA_ADDR_H       ; enable auto increment by 2, bank 0.
+        lda  #1
+        sta  cx16.VERA_ADDR_L       ; start at (1,0)
+        stz  cx16.VERA_ADDR_M
+_lx     ldx  #0                     ; modified
+_la     lda  #0                     ; modified
+-       sta  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA0
+        dex
+        bne  -
+        dey
+        beq  +
+        lda  #1
+        sta  cx16.VERA_ADDR_L
+        inc  cx16.VERA_ADDR_M       ; next line
+        bra  _lx
++       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) {
+    c64.CHROUT(color_to_charcode[txtcol & 15])
+}
+
+sub color2 (ubyte txtcol, ubyte bgcol) {
+    c64.CHROUT(color_to_charcode[bgcol & 15])
+    c64.CHROUT(1)       ; switch fg and bg colors
+    c64.CHROUT(color_to_charcode[txtcol & 15])
+}
+
+sub lowercase() {
+    cx16.screen_set_charset(3, 0)  ; lowercase charset
+}
+
+sub uppercase() {
+    cx16.screen_set_charset(2, 0)  ; uppercase charset
+}
+
+asmsub  scroll_left  (ubyte dummy @ Pc) clobbers(A, Y)  {
+	; ---- scroll the whole screen 1 character to the left
+	;      contents of the rightmost column are unchanged, you should clear/refill this yourself
+	;      Carry flag is a dummy on the cx16
+	%asm {{
+	    phx
+	    jsr  c64.SCREEN
+	    dex
+	    stx  _lx+1
+        dey
+        sty  P8ZP_SCRATCH_B1    ; number of rows to scroll
+
+_nextline
+        stz  cx16.VERA_CTRL     ; data port 0: source column
+        lda  #%00010000         ; auto increment 1
+        sta  cx16.VERA_ADDR_H
+        lda  #2
+        sta  cx16.VERA_ADDR_L   ; begin in column 1
+        ldy  P8ZP_SCRATCH_B1
+        sty  cx16.VERA_ADDR_M
+        lda  #1
+        sta  cx16.VERA_CTRL     ; data port 1: destination column
+        lda  #%00010000         ; auto increment 1
+        sta  cx16.VERA_ADDR_H
+        stz  cx16.VERA_ADDR_L
+        sty  cx16.VERA_ADDR_M
+
+_lx     ldx  #0                ; modified
+-       lda  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA1    ; copy char
+        lda  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA1    ; copy color
+        dex
+        bne  -
+        dec  P8ZP_SCRATCH_B1
+        bpl  _nextline
+
+        lda  #0
+        sta  cx16.VERA_CTRL
+	    plx
+	    rts
+	}}
+}
+
+asmsub  scroll_right  (ubyte dummy @ Pc) clobbers(A)  {
+	; ---- scroll the whole screen 1 character to the right
+	;      contents of the leftmost column are unchanged, you should clear/refill this yourself
+	;      Carry flag is a dummy on the cx16
+	%asm {{
+	    phx
+	    jsr  c64.SCREEN
+	    dex
+	    stx  _lx+1
+	    txa
+	    asl  a
+	    dea
+	    sta  _rcol+1
+	    ina
+	    ina
+	    sta  _rcol2+1
+        dey
+        sty  P8ZP_SCRATCH_B1    ; number of rows to scroll
+
+_nextline
+        stz  cx16.VERA_CTRL     ; data port 0: source column
+        lda  #%00011000         ; auto decrement 1
+        sta  cx16.VERA_ADDR_H
+_rcol   lda  #79*2-1             ; modified
+        sta  cx16.VERA_ADDR_L   ; begin in rightmost column minus one
+        ldy  P8ZP_SCRATCH_B1
+        sty  cx16.VERA_ADDR_M
+        lda  #1
+        sta  cx16.VERA_CTRL     ; data port 1: destination column
+        lda  #%00011000         ; auto decrement 1
+        sta  cx16.VERA_ADDR_H
+_rcol2  lda  #79*2+1            ; modified
+        sta  cx16.VERA_ADDR_L
+        sty  cx16.VERA_ADDR_M
+
+_lx     ldx  #0                 ; modified
+-       lda  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA1    ; copy char
+        lda  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA1    ; copy color
+        dex
+        bne  -
+        dec  P8ZP_SCRATCH_B1
+        bpl  _nextline
+
+        lda  #0
+        sta  cx16.VERA_CTRL
+	    plx
+	    rts
+	}}
+}
+
+asmsub  scroll_up  (ubyte dummy @ Pc) clobbers(A, Y)  {
+	; ---- scroll the whole screen 1 character up
+	;      contents of the bottom row are unchanged, you should refill/clear this yourself
+	;      Carry flag is a dummy on the cx16
+	%asm {{
+	    phx
+	    jsr  c64.SCREEN
+	    stx  _nextline+1
+	    dey
+        sty  P8ZP_SCRATCH_B1
+        stz  cx16.VERA_CTRL         ; data port 0 is source
+        lda  #1
+        sta  cx16.VERA_ADDR_M       ; start at second line
+        stz  cx16.VERA_ADDR_L
+        lda  #%00010000
+        sta  cx16.VERA_ADDR_H       ; enable auto increment by 1, bank 0.
+
+        lda  #1
+        sta  cx16.VERA_CTRL         ; data port 1 is destination
+        stz  cx16.VERA_ADDR_M       ; start at top line
+        stz  cx16.VERA_ADDR_L
+        lda  #%00010000
+        sta  cx16.VERA_ADDR_H       ; enable auto increment by 1, bank 0.
+
+_nextline
+        ldx  #80        ; modified
+-       lda  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA1        ; copy char
+        lda  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA1        ; copy color
+        dex
+        bne  -
+        dec  P8ZP_SCRATCH_B1
+        beq  +
+        stz  cx16.VERA_CTRL         ; data port 0
+        stz  cx16.VERA_ADDR_L
+        inc  cx16.VERA_ADDR_M
+        lda  #1
+        sta  cx16.VERA_CTRL         ; data port 1
+        stz  cx16.VERA_ADDR_L
+        inc  cx16.VERA_ADDR_M
+        bra  _nextline
+
++       lda  #0
+        sta  cx16.VERA_CTRL
+	    plx
+	    rts
+	}}
+}
+
+asmsub  scroll_down  (ubyte dummy @ Pc) clobbers(A, Y)  {
+	; ---- scroll the whole screen 1 character down
+	;      contents of the top row are unchanged, you should refill/clear this yourself
+	;      Carry flag is a dummy on the cx16
+	%asm {{
+	    phx
+	    jsr  c64.SCREEN
+	    stx  _nextline+1
+	    dey
+        sty  P8ZP_SCRATCH_B1
+        stz  cx16.VERA_CTRL         ; data port 0 is source
+        dey
+        sty  cx16.VERA_ADDR_M       ; start at line before bottom line
+        stz  cx16.VERA_ADDR_L
+        lda  #%00010000
+        sta  cx16.VERA_ADDR_H       ; enable auto increment by 1, bank 0.
+
+        lda  #1
+        sta  cx16.VERA_CTRL         ; data port 1 is destination
+        iny
+        sty  cx16.VERA_ADDR_M       ; start at bottom line
+        stz  cx16.VERA_ADDR_L
+        lda  #%00010000
+        sta  cx16.VERA_ADDR_H       ; enable auto increment by 1, bank 0.
+
+_nextline
+        ldx  #80        ; modified
+-       lda  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA1        ; copy char
+        lda  cx16.VERA_DATA0
+        sta  cx16.VERA_DATA1        ; copy color
+        dex
+        bne  -
+        dec  P8ZP_SCRATCH_B1
+        beq  +
+        stz  cx16.VERA_CTRL         ; data port 0
+        stz  cx16.VERA_ADDR_L
+        dec  cx16.VERA_ADDR_M
+        lda  #1
+        sta  cx16.VERA_CTRL         ; data port 1
+        stz  cx16.VERA_ADDR_L
+        dec  cx16.VERA_ADDR_M
+        bra  _nextline
+
++       lda  #0
+        sta  cx16.VERA_CTRL
+	    plx
+	    rts
+	}}
+}
+
+romsub $FFD2 = chrout(ubyte char @ A)    ; for consistency. You can also use c64.CHROUT directly ofcourse.
+
+asmsub  print (str text @ AY) clobbers(A,Y)  {
+	; ---- print null terminated string from A/Y
+	; note: the compiler contains an optimization that will replace
+	;       a call to this subroutine with a string argument of just one char,
+	;       by just one call to c64.CHROUT of that single char.
+	%asm {{
+		sta  P8ZP_SCRATCH_B1
+		sty  P8ZP_SCRATCH_REG
+		ldy  #0
+-		lda  (P8ZP_SCRATCH_B1),y
+		beq  +
+		jsr  c64.CHROUT
+		iny
+		bne  -
++		rts
+	}}
+}
+
+asmsub  print_ub0  (ubyte value @ A) clobbers(A,Y)  {
+	; ---- print the ubyte in A in decimal form, with left padding 0s (3 positions total)
+	%asm {{
+		phx
+		jsr  conv.ubyte2decimal
+		pha
+		tya
+		jsr  c64.CHROUT
+		pla
+		jsr  c64.CHROUT
+		txa
+		jsr  c64.CHROUT
+		plx
+		rts
+	}}
+}
+
+asmsub  print_ub  (ubyte value @ A) clobbers(A,Y)  {
+	; ---- print the ubyte in A in decimal form, without left padding 0s
+	%asm {{
+		phx
+		jsr  conv.ubyte2decimal
+_print_byte_digits
+		pha
+		cpy  #'0'
+		beq  +
+		tya
+		jsr  c64.CHROUT
+		pla
+		jsr  c64.CHROUT
+		jmp  _ones
++       pla
+        cmp  #'0'
+        beq  _ones
+        jsr  c64.CHROUT
+_ones   txa
+		jsr  c64.CHROUT
+		plx
+		rts
+	}}
+}
+
+asmsub  print_b  (byte value @ A) clobbers(A,Y)  {
+	; ---- print the byte in A in decimal form, without left padding 0s
+	%asm {{
+		phx
+		pha
+		cmp  #0
+		bpl  +
+		lda  #'-'
+		jsr  c64.CHROUT
++		pla
+		jsr  conv.byte2decimal
+		jmp  print_ub._print_byte_digits
+	}}
+}
+
+asmsub  print_ubhex  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
+	; ---- print the ubyte in A in hex form (if Carry is set, a radix prefix '$' is printed as well)
+	%asm {{
+		phx
+		bcc  +
+		pha
+		lda  #'$'
+		jsr  c64.CHROUT
+		pla
++		jsr  conv.ubyte2hex
+		jsr  c64.CHROUT
+		tya
+		jsr  c64.CHROUT
+		plx
+		rts
+	}}
+}
+
+asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
+	; ---- print the ubyte in A in binary form (if Carry is set, a radix prefix '%' is printed as well)
+	%asm {{
+		phx
+		sta  P8ZP_SCRATCH_B1
+		bcc  +
+		lda  #'%'
+		jsr  c64.CHROUT
++		ldy  #8
+-		lda  #'0'
+		asl  P8ZP_SCRATCH_B1
+		bcc  +
+		lda  #'1'
++		jsr  c64.CHROUT
+		dey
+		bne  -
+		plx
+		rts
+	}}
+}
+
+asmsub  print_uwbin  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
+	; ---- print the uword in A/Y in binary form (if Carry is set, a radix prefix '%' is printed as well)
+	%asm {{
+		pha
+		tya
+		jsr  print_ubbin
+		pla
+		clc
+		jmp  print_ubbin
+	}}
+}
+
+asmsub  print_uwhex  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
+	; ---- print the uword in A/Y in hexadecimal form (4 digits)
+	;      (if Carry is set, a radix prefix '$' is printed as well)
+	%asm {{
+		pha
+		tya
+		jsr  print_ubhex
+		pla
+		clc
+		jmp  print_ubhex
+	}}
+}
+
+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 {{
+	    phx
+		jsr  conv.uword2decimal
+		ldy  #0
+-		lda  conv.uword2decimal.decTenThousands,y
+        beq  +
+		jsr  c64.CHROUT
+		iny
+		bne  -
++		plx
+		rts
+	}}
+}
+
+asmsub  print_uw  (uword value @ AY) clobbers(A,Y)  {
+	; ---- print the uword in A/Y in decimal form, without left padding 0s
+	%asm {{
+	    phx
+		jsr  conv.uword2decimal
+		plx
+		ldy  #0
+-		lda  conv.uword2decimal.decTenThousands,y
+		beq  _allzero
+		cmp  #'0'
+		bne  _gotdigit
+		iny
+		bne  -
+
+_gotdigit
+		jsr  c64.CHROUT
+		iny
+		lda  conv.uword2decimal.decTenThousands,y
+		bne  _gotdigit
+		rts
+_allzero
+        lda  #'0'
+        jmp  c64.CHROUT
+	}}
+}
+
+asmsub  print_w  (word value @ AY) clobbers(A,Y)  {
+	; ---- print the (signed) word in A/Y in decimal form, without left padding 0's
+	%asm {{
+		cpy  #0
+		bpl  +
+		pha
+		lda  #'-'
+		jsr  c64.CHROUT
+		tya
+		eor  #255
+		tay
+		pla
+		eor  #255
+		clc
+		adc  #1
+		bcc  +
+		iny
++		jmp  print_uw
+	}}
+}
+
+asmsub  input_chars  (uword buffer @ AY) clobbers(A) -> ubyte @ Y  {
+	; ---- Input a string (max. 80 chars) from the keyboard. Returns length in Y. (string is terminated with a 0 byte as well)
+	;      It assumes the keyboard is selected as I/O channel!
+
+	%asm {{
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		ldy  #0				; char counter = 0
+-		jsr  c64.CHRIN
+		cmp  #$0d			; return (ascii 13) pressed?
+		beq  +				; yes, end.
+		sta  (P8ZP_SCRATCH_W1),y	; else store char in buffer
+		iny
+		bne  -
++		lda  #0
+		sta  (P8ZP_SCRATCH_W1),y	; finish string with 0 byte
+		rts
+
+	}}
+}
+
+asmsub  setchr  (ubyte col @X, ubyte row @Y, ubyte character @A) clobbers(A)  {
+	; ---- sets the character in the screen matrix at the given position
+	%asm {{
+		pha
+		txa
+		asl  a
+        stz  cx16.VERA_CTRL
+		stz  cx16.VERA_ADDR_H
+		sta  cx16.VERA_ADDR_L
+		sty  cx16.VERA_ADDR_M
+		pla
+        sta  cx16.VERA_DATA0
+		rts
+	}}
+}
+
+asmsub  getchr  (ubyte col @A, ubyte row @Y) -> ubyte @ A {
+	; ---- get the character in the screen matrix at the given location
+	%asm  {{
+		asl  a
+        stz  cx16.VERA_CTRL
+		stz  cx16.VERA_ADDR_H
+		sta  cx16.VERA_ADDR_L
+		sty  cx16.VERA_ADDR_M
+        lda  cx16.VERA_DATA0
+		rts
+	}}
+}
+
+asmsub  setclr  (ubyte col @X, ubyte row @Y, ubyte color @A) clobbers(A)  {
+	; ---- set the color in A on the screen matrix at the given position
+	%asm {{
+		pha
+		txa
+		asl  a
+		ina
+        stz  cx16.VERA_CTRL
+		stz  cx16.VERA_ADDR_H
+		sta  cx16.VERA_ADDR_L
+		sty  cx16.VERA_ADDR_M
+		pla
+        sta  cx16.VERA_DATA0
+		rts
+	}}
+}
+
+asmsub  getclr  (ubyte col @A, ubyte row @Y) -> ubyte @ A {
+	; ---- get the color in the screen color matrix at the given location
+	%asm  {{
+		asl  a
+		ina
+        stz  cx16.VERA_CTRL
+		stz  cx16.VERA_ADDR_H
+		sta  cx16.VERA_ADDR_L
+		sty  cx16.VERA_ADDR_M
+        lda  cx16.VERA_DATA0
+		rts
+	}}
+}
+
+sub  setcc  (ubyte column, ubyte row, ubyte char, ubyte charcolor)  {
+	; ---- set char+color at the given position on the screen
+	%asm {{
+	    phx
+		lda  column
+		asl  a
+		tax
+		ldy  row
+		lda  charcolor
+		and  #$0f
+	    sta  P8ZP_SCRATCH_B1
+        stz  cx16.VERA_CTRL
+        stz  cx16.VERA_ADDR_H
+        stx  cx16.VERA_ADDR_L
+        sty  cx16.VERA_ADDR_M
+        lda  char
+        sta  cx16.VERA_DATA0
+        inx
+        stz  cx16.VERA_ADDR_H
+        stx  cx16.VERA_ADDR_L
+        sty  cx16.VERA_ADDR_M
+        lda  cx16.VERA_DATA0
+        and  #$f0
+        ora  P8ZP_SCRATCH_B1
+        sta  cx16.VERA_DATA0
+		plx
+		rts
+    }}
+}
+
+asmsub  plot  (ubyte col @ Y, ubyte row @ A) clobbers(A) {
+	; ---- safe wrapper around PLOT kernel routine, to save the X register.
+	%asm  {{
+		phx
+		tax
+		clc
+		jsr  c64.PLOT
+		plx
+		rts
+	}}
+}
+
+asmsub width() clobbers(X,Y) -> ubyte @A {
+    ; -- returns the text screen width (number of columns)
+    %asm {{
+        jsr  c64.SCREEN
+        txa
+        rts
+    }}
+}
+
+asmsub height() clobbers(X, Y) -> ubyte @A {
+    ; -- returns the text screen height (number of rows)
+    %asm {{
+        jsr  c64.SCREEN
+        tya
+        rts
+    }}
+}
+
+}

compiler/res/prog8lib/cx16lib.p8

@@ -1,221 +0,0 @@
-; Prog8 definitions for the CommanderX16
-; Including memory registers, I/O registers, Basic and Kernal subroutines.
-;
-; Written by Irmen de Jong (irmen@razorvine.net) - license: GNU GPL 3.0
-;
-; indent format: TABS, size=8
-
-
-c64 {
-
-; ---- kernal routines, these are the same as on the Commodore-64 (hence the same block name) ----
-
-; STROUT --> use screen.print
-; 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 $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 $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 $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 $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 $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 $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
-
-}
-
-cx16 {
-
-; ---- Commander X-16 additions on top of C64 kernal routines ----
-; spelling of the names is taken from the Commander X-16 rom sources
-
-; the sixteen virtual 16-bit registers
-    &uword r0  = $02
-    &uword r1  = $04
-    &uword r2  = $06
-    &uword r3  = $08
-    &uword r4  = $0a
-    &uword r5  = $0c
-    &uword r6  = $0e
-    &uword r7  = $10
-    &uword r8  = $12
-    &uword r9  = $14
-    &uword r10 = $16
-    &uword r11 = $18
-    &uword r12 = $1a
-    &uword r13 = $1c
-    &uword r14 = $1e
-    &uword r15 = $20
-
-; 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
-; VERA_PSG_BASE     = $1F9C0
-; VERA_PALETTE_BASE = $1FA00
-; VERA_SPRITES_BASE = $1FC00
-
-
-; 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()
-
-; 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 $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
-
-; 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
-
-; 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
-
-; 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 $fecf = entropy_get() -> ubyte @A, ubyte @X, ubyte @Y
-romsub $fecc = monitor()
-
-
-
-; ---- end of kernal routines ----
-
-asmsub init_system()  {
-    ; Initializes the machine to a sane starting state.
-    ; Called automatically by the loader program logic.
-    %asm {{
-        sei
-        cld
-        stz  $00
-        stz  $01
-        jsr  c64.IOINIT
-        jsr  c64.RESTOR
-        jsr  c64.CINT
-        lda  #0
-        tax
-        tay
-        clc
-        clv
-        cli
-        rts
-    }}
-}
-
-}

compiler/res/prog8lib/cx16textio.p8

@@ -1,314 +0,0 @@
-; Prog8 definitions for the Text I/O and Screen routines for the CommanderX16
-;
-; Written by Irmen de Jong (irmen@razorvine.net) - license: GNU GPL 3.0
-;
-; indent format: TABS, size=8
-
-
-%import cx16lib
-%import conv
-
-
-txt {
-
-sub  clear_screen() {
-    c64.CHROUT(147)         ; clear screen (spaces)
-}
-
-
-asmsub  fill_screen (ubyte char @ A, ubyte txtcolor @ Y) clobbers(A)  {
-	; ---- fill the character screen with the given fill character and character color.
-	%asm {{
-        sta  P8ZP_SCRATCH_W1        ; fillchar
-        sty  P8ZP_SCRATCH_W1+1      ; textcolor
-        phx
-        jsr  c64.SCREEN             ; get dimensions in X/Y
-        dex
-        dey
-        txa
-        asl  a
-        adc  #1
-        sta  P8ZP_SCRATCH_B1
--       ldx  P8ZP_SCRATCH_B1
--       stz  cx16.VERA_ADDR_H
-        stx  cx16.VERA_ADDR_L
-        sty  cx16.VERA_ADDR_M
-        lda  cx16.VERA_DATA0
-        and  #$f0
-        ora  P8ZP_SCRATCH_W1+1
-        sta  cx16.VERA_DATA0
-        dex
-        stz  cx16.VERA_ADDR_H
-        stx  cx16.VERA_ADDR_L
-        sty  cx16.VERA_ADDR_M
-        lda  P8ZP_SCRATCH_W1
-        sta  cx16.VERA_DATA0
-        dex
-        cpx  #255
-        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,$9b]
-
-sub color (ubyte txtcol) {
-    c64.CHROUT(color_to_charcode[txtcol & 15])
-}
-
-sub color2 (ubyte txtcol, ubyte bgcol) {
-    c64.CHROUT(color_to_charcode[bgcol & 15])
-    c64.CHROUT(1)       ; switch fg and bg colors
-    c64.CHROUT(color_to_charcode[txtcol & 15])
-}
-
-asmsub  print (str text @ AY) clobbers(A,Y)  {
-	; ---- print null terminated string from A/Y
-	; note: the compiler contains an optimization that will replace
-	;       a call to this subroutine with a string argument of just one char,
-	;       by just one call to c64.CHROUT of that single char.
-	%asm {{
-		sta  P8ZP_SCRATCH_B1
-		sty  P8ZP_SCRATCH_REG
-		ldy  #0
--		lda  (P8ZP_SCRATCH_B1),y
-		beq  +
-		jsr  c64.CHROUT
-		iny
-		bne  -
-+		rts
-	}}
-}
-
-asmsub  print_ub0  (ubyte value @ A) clobbers(A,Y)  {
-	; ---- print the ubyte in A in decimal form, with left padding 0s (3 positions total)
-	%asm {{
-		phx
-		jsr  conv.ubyte2decimal
-		pha
-		tya
-		jsr  c64.CHROUT
-		pla
-		jsr  c64.CHROUT
-		txa
-		jsr  c64.CHROUT
-		plx
-		rts
-	}}
-}
-
-asmsub  print_ub  (ubyte value @ A) clobbers(A,Y)  {
-	; ---- print the ubyte in A in decimal form, without left padding 0s
-	%asm {{
-		phx
-		jsr  conv.ubyte2decimal
-_print_byte_digits
-		pha
-		cpy  #'0'
-		beq  +
-		tya
-		jsr  c64.CHROUT
-		pla
-		jsr  c64.CHROUT
-		jmp  _ones
-+       pla
-        cmp  #'0'
-        beq  _ones
-        jsr  c64.CHROUT
-_ones   txa
-		jsr  c64.CHROUT
-		plx
-		rts
-	}}
-}
-
-asmsub  print_b  (byte value @ A) clobbers(A,Y)  {
-	; ---- print the byte in A in decimal form, without left padding 0s
-	%asm {{
-		phx
-		pha
-		cmp  #0
-		bpl  +
-		lda  #'-'
-		jsr  c64.CHROUT
-+		pla
-		jsr  conv.byte2decimal
-		jmp  print_ub._print_byte_digits
-	}}
-}
-
-asmsub  print_ubhex  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
-	; ---- print the ubyte in A in hex form (if Carry is set, a radix prefix '$' is printed as well)
-	%asm {{
-		phx
-		bcc  +
-		pha
-		lda  #'$'
-		jsr  c64.CHROUT
-		pla
-+		jsr  conv.ubyte2hex
-		jsr  c64.CHROUT
-		tya
-		jsr  c64.CHROUT
-		plx
-		rts
-	}}
-}
-
-asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
-	; ---- print the ubyte in A in binary form (if Carry is set, a radix prefix '%' is printed as well)
-	%asm {{
-		phx
-		sta  P8ZP_SCRATCH_B1
-		bcc  +
-		lda  #'%'
-		jsr  c64.CHROUT
-+		ldy  #8
--		lda  #'0'
-		asl  P8ZP_SCRATCH_B1
-		bcc  +
-		lda  #'1'
-+		jsr  c64.CHROUT
-		dey
-		bne  -
-		plx
-		rts
-	}}
-}
-
-asmsub  print_uwbin  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
-	; ---- print the uword in A/Y in binary form (if Carry is set, a radix prefix '%' is printed as well)
-	%asm {{
-		pha
-		tya
-		jsr  print_ubbin
-		pla
-		clc
-		jmp  print_ubbin
-	}}
-}
-
-asmsub  print_uwhex  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
-	; ---- print the uword in A/Y in hexadecimal form (4 digits)
-	;      (if Carry is set, a radix prefix '$' is printed as well)
-	%asm {{
-		pha
-		tya
-		jsr  print_ubhex
-		pla
-		clc
-		jmp  print_ubhex
-	}}
-}
-
-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 {{
-	    phx
-		jsr  conv.uword2decimal
-		ldy  #0
--		lda  conv.uword2decimal.decTenThousands,y
-        beq  +
-		jsr  c64.CHROUT
-		iny
-		bne  -
-+		plx
-		rts
-	}}
-}
-
-asmsub  print_uw  (uword value @ AY) clobbers(A,Y)  {
-	; ---- print the uword in A/Y in decimal form, without left padding 0s
-	%asm {{
-	    phx
-		jsr  conv.uword2decimal
-		plx
-		ldy  #0
--		lda  conv.uword2decimal.decTenThousands,y
-		beq  _allzero
-		cmp  #'0'
-		bne  _gotdigit
-		iny
-		bne  -
-
-_gotdigit
-		jsr  c64.CHROUT
-		iny
-		lda  conv.uword2decimal.decTenThousands,y
-		bne  _gotdigit
-		rts
-_allzero
-        lda  #'0'
-        jmp  c64.CHROUT
-	}}
-}
-
-asmsub  print_w  (word value @ AY) clobbers(A,Y)  {
-	; ---- print the (signed) word in A/Y in decimal form, without left padding 0's
-	%asm {{
-		cpy  #0
-		bpl  +
-		pha
-		lda  #'-'
-		jsr  c64.CHROUT
-		tya
-		eor  #255
-		tay
-		pla
-		eor  #255
-		clc
-		adc  #1
-		bcc  +
-		iny
-+		jmp  print_uw
-	}}
-}
-
-; TODO implement the "missing" txtio subroutines
-
-sub  setcc  (ubyte column, ubyte row, ubyte char, ubyte charcolor)  {
-	; ---- set char+color at the given position on the screen
-	%asm {{
-	    phx
-		lda  column
-		asl  a
-		tax
-		ldy  row
-		lda  charcolor
-		and  #$0f
-	    sta  P8ZP_SCRATCH_B1
-        stz  cx16.VERA_ADDR_H
-        stx  cx16.VERA_ADDR_L
-        sty  cx16.VERA_ADDR_M
-        lda  char
-        sta  cx16.VERA_DATA0
-        inx
-        stz  cx16.VERA_ADDR_H
-        stx  cx16.VERA_ADDR_L
-        sty  cx16.VERA_ADDR_M
-        lda  cx16.VERA_DATA0
-        and  #$f0
-        ora  P8ZP_SCRATCH_B1
-        sta  cx16.VERA_DATA0
-		plx
-		rts
-    }}
-}
-
-asmsub  plot  (ubyte col @ Y, ubyte row @ A) clobbers(A) {
-	; ---- safe wrapper around PLOT kernel routine, to save the X register.
-	%asm  {{
-		phx
-		tax
-		clc
-		jsr  c64.PLOT
-		plx
-		rts
-	}}
-}
-
-}

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
@@ -27,11 +31,11 @@ _enterloop	lsr  P8ZP_SCRATCH_REG
 		.pend
 
 
-multiply_bytes_16	.proc
+multiply_bytes_into_word	.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
+stack_mul_word_25	.proc
-		; W + W*8 + W*16
+		; 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
+stack_mul_byte_40	.proc
-		; (X*4 + X)*8
 		lda  P8ESTACK_LO+1,x
-		asl  a
+		and  #7
-		asl  a
+		tay
-                clc
+		lda  mul_byte_40._forties,y
-		adc  P8ESTACK_LO+1,x
-		asl  a
-		asl  a
-		asl  a
 		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
+4.3

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

@@ -30,7 +30,7 @@ private fun ParserRuleContext.toPosition() : Position {
     val customTokensource = this.start.tokenSource as? CustomLexer
     val filename =
             when {
-                customTokensource!=null -> customTokensource.modulePath.fileName.toString()
+                customTokensource!=null -> customTokensource.modulePath.toString()
                 start.tokenSource.sourceName == IntStream.UNKNOWN_SOURCE_NAME -> "@internal@"
                 else -> File(start.inputStream.sourceName).name
             }
@@ -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,
+    AsmSubroutineParameter(vardecl.varname.text, datatype, registerorpair, statusregister, toPosition())
-            !it.stack?.text.isNullOrEmpty(), toPosition())
 }
 
 private fun prog8Parser.Functioncall_stmtContext.toAst(): Statement {
@@ -398,7 +397,7 @@ private fun prog8Parser.DirectiveargContext.toAst() : DirectiveArg {
 }
 
 private fun prog8Parser.IntegerliteralContext.toAst(): NumericLiteral {
-    fun makeLiteral(text: String, radix: Int, forceWord: Boolean): NumericLiteral {
+    fun makeLiteral(text: String, radix: Int): NumericLiteral {
         val integer: Int
         var datatype = DataType.UBYTE
         when (radix) {
@@ -436,14 +435,14 @@ private fun prog8Parser.IntegerliteralContext.toAst(): NumericLiteral {
             }
             else -> throw FatalAstException("invalid radix")
         }
-        return NumericLiteral(integer, if (forceWord) DataType.UWORD else datatype)
+        return NumericLiteral(integer, datatype)
     }
     val terminal: TerminalNode = children[0] as TerminalNode
     val integerPart = this.intpart.text
     return when (terminal.symbol.type) {
-        prog8Parser.DEC_INTEGER -> makeLiteral(integerPart, 10, wordsuffix()!=null)
+        prog8Parser.DEC_INTEGER -> makeLiteral(integerPart, 10)
-        prog8Parser.HEX_INTEGER -> makeLiteral(integerPart.substring(1), 16, wordsuffix()!=null)
+        prog8Parser.HEX_INTEGER -> makeLiteral(integerPart.substring(1), 16)
-        prog8Parser.BIN_INTEGER -> makeLiteral(integerPart.substring(1), 2, wordsuffix()!=null)
+        prog8Parser.BIN_INTEGER -> makeLiteral(integerPart.substring(1), 2)
         else -> throw FatalAstException(terminal.text)
     }
 }
@@ -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
+            FLOAT -> CompilationTarget.instance.machine.FLOAT_MEM_SIZE
-            in PassByReferenceDatatypes -> CompilationTarget.machine.POINTER_MEM_SIZE
+            in PassByReferenceDatatypes -> CompilationTarget.instance.machine.POINTER_MEM_SIZE
             else -> -9999999
         }
     }
@@ -165,6 +165,7 @@ object ParentSentinel : Node {
 
 data class Position(val file: String, val line: Int, val startCol: Int, val endCol: Int) {
     override fun toString(): String = "[$file: line $line col ${startCol+1}-${endCol+1}]"
+    fun toClickableStr(): String = "$file:$line:$startCol:"
 
     companion object {
         val DUMMY = Position("<dummy>", 0, 0, 0)

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

@@ -24,7 +24,7 @@ class ErrorReporter {
                 MessageSeverity.ERROR -> System.err.print("\u001b[91m")  // bright red
                 MessageSeverity.WARNING -> System.err.print("\u001b[93m")  // bright yellow
             }
-            val msg = "${it.position} ${it.severity} ${it.message}".trim()
+            val msg = "${it.position.toClickableStr()} ${it.severity} ${it.message}".trim()
             if(msg !in alreadyReportedMessages) {
                 System.err.println(msg)
                 alreadyReportedMessages.add(msg)

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

@@ -7,11 +7,11 @@ open class FatalAstException (override var message: String) : Exception(message)
 open class AstException (override var message: String) : Exception(message)
 
 open class SyntaxError(override var message: String, val position: Position) : AstException(message) {
-    override fun toString() = "$position Syntax error: $message"
+    override fun toString() = "${position.toClickableStr()} Syntax error: $message"
 }
 
 class ExpressionError(message: String, val position: Position) : AstException(message) {
-    override fun toString() = "$position Error: $message"
+    override fun toString() = "${position.toClickableStr()} Error: $message"
 }
 
 class UndefinedSymbolError(symbol: IdentifierReference)

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()
+            fromVal = CompilationTarget.instance.encodeString(fromString.value, fromString.altEncoding)[0].toInt()
-            toVal = CompilationTarget.encodeString(toString.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
@@ -744,7 +744,8 @@ data class IdentifierReference(val nameInSource: List<String>, override val posi
     override fun accept(visitor: IAstVisitor) = visitor.visit(this)
     override fun accept(visitor: AstWalker, parent: Node)= visitor.visit(this, parent)
 
-    override fun referencesIdentifiers(vararg name: String): Boolean = nameInSource.last() in name
+    override fun referencesIdentifiers(vararg name: String): Boolean =
+            nameInSource.size==name.size && nameInSource.toTypedArray().contentEquals(name)
 
     override fun inferType(program: Program): InferredTypes.InferredType {
         return when (val targetStmt = targetStatement(program.namespace)) {

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) {
+        fun err(msg: String, position: Position?=null) = errors.err(msg, position ?: decl.position)
-            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
+        // FLOATS enabled?
-        if(!compilerOptions.floats && decl.datatype in setOf(DataType.FLOAT, DataType.ARRAY_F) && decl.type!= VarDeclType.MEMORY) {
+        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) {
@@ -558,12 +559,14 @@ 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}")
+                        if(decl.type==VarDeclType.CONST) {
+                            err("const declaration needs a compile-time constant initializer value, or range")
                             super.visit(decl)
                             return
                         }
                     }
                 }
+            }
             VarDeclType.MEMORY -> {
                 if(decl.arraysize!=null) {
                     val arraySize = decl.arraysize!!.constIndex() ?: 1
@@ -581,13 +584,13 @@ internal class AstChecker(private val program: Program,
                     }
                 }
 
-                if(decl.value !is NumericLiteralValue) {
+                if(decl.value is NumericLiteralValue) {
-                    err("value of memory var decl is not a numeric literal (it is a ${decl.value!!.javaClass.simpleName}).", decl.value?.position)
-                } else {
                     val value = decl.value as NumericLiteralValue
                     if (value.type !in IntegerDatatypes || value.number.toInt() < 0 || value.number.toInt() > 65535) {
                         err("memory address must be valid integer 0..\$ffff", decl.value?.position)
                     }
+                } else {
+                    err("value of memory mapped variable can only be a number, perhaps you meant to use an address pointer type instead?", decl.value?.position)
                 }
             }
         }
@@ -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)
@@ -731,6 +742,11 @@ internal class AstChecker(private val program: Program,
             checkValueTypeAndRangeArray(array.type.typeOrElse(DataType.STRUCT), null, arrayspec, array)
         }
 
+        if(!array.value.all { it is NumericLiteralValue || it is AddressOf || it is StringLiteralValue }) {
+            // TODO for now, array literals have to consist of all compile time constant values...
+            errors.err("array literal doesn't consist of only compile time constant values", array.position)
+        }
+
         super.visit(array)
     }
 
@@ -1087,7 +1103,7 @@ internal class AstChecker(private val program: Program,
         }
 
         if(value.type.isUnknown)
-            return err("attempt to check values of array with as yet unknown datatype")
+            return false
 
         when (targetDt) {
             DataType.STR -> return err("string value expected")
@@ -1156,7 +1172,7 @@ 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]
@@ -34,13 +34,23 @@ internal class AstIdentifiersChecker(private val program: Program, private val e
         super.visit(block)
     }
 
+    override fun visit(directive: Directive) {
+        if(directive.directive=="%target") {
+            val compatibleTarget = directive.args.single().name
+            if (compatibleTarget != CompilationTarget.instance.name)
+                errors.err("module's compilation target ($compatibleTarget) differs from active target (${CompilationTarget.instance.name})", directive.position)
+        }
+
+        super.visit(directive)
+    }
+
     override fun visit(decl: VarDecl) {
         decl.datatypeErrors.forEach { errors.err(it.message, it.position) }
 
         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 +84,7 @@ 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
@@ -88,14 +98,6 @@ internal class AstIdentifiersChecker(private val program: Program, private val e
             if (existing != null && existing !== subroutine)
                 nameError(subroutine.name, subroutine.position, existing)
 
-            // does the parameter redefine a variable declared elsewhere?
-            for(param in subroutine.parameters) {
-                val existingVar = subroutine.lookup(listOf(param.name), subroutine)
-                if (existingVar != null && existingVar.parent !== subroutine) {
-                    nameError(param.name, param.position, existingVar)
-                }
-            }
-
             // check that there are no local variables, labels, or other subs that redefine the subroutine's parameters
             val symbolsInSub = subroutine.allDefinedSymbols()
             val namesInSub = symbolsInSub.map{ it.first }.toSet()
@@ -119,7 +121,7 @@ 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

@@ -10,43 +10,6 @@ import prog8.ast.statements.*
 internal class AstVariousTransforms(private val program: Program) : AstWalker() {
     private val noModifications = emptyList<IAstModification>()
 
-    override fun after(functionCallStatement: FunctionCallStatement, parent: Node): Iterable<IAstModification> {
-        if(functionCallStatement.target.nameInSource == listOf("swap")) {
-            // 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:
-            // - declare a temp variable of the same datatype
-            // - temp = x, x = y, y= temp
-            val first = functionCallStatement.args[0]
-            val second = functionCallStatement.args[1]
-            if(first !is IdentifierReference && second !is IdentifierReference) {
-                val dt = first.inferType(program).typeOrElse(DataType.STRUCT)
-                val tempname = "prog8_swaptmp_${functionCallStatement.hashCode()}"
-                val tempvardecl = VarDecl(VarDeclType.VAR, dt, ZeropageWish.DONTCARE, null, tempname, null, null, isArray = false, autogeneratedDontRemove = true, position = first.position)
-                val tempvar = IdentifierReference(listOf(tempname), first.position)
-                val assignTemp = Assignment(
-                        AssignTarget(tempvar, null, null, first.position),
-                        first,
-                        first.position
-                )
-                val assignFirst = Assignment(
-                        AssignTarget.fromExpr(first),
-                        second,
-                        first.position
-                )
-                val assignSecond = Assignment(
-                        AssignTarget.fromExpr(second),
-                        tempvar,
-                        first.position
-                )
-                val scope = AnonymousScope(mutableListOf(tempvardecl, assignTemp, assignFirst, assignSecond), first.position)
-                return listOf(IAstModification.ReplaceNode(functionCallStatement, scope, parent))
-            }
-        }
-
-        return noModifications
-    }
-
     override fun before(decl: VarDecl, parent: Node): Iterable<IAstModification> {
         // is it a struct variable? then define all its struct members as mangled names,
         //    and include the original decl as well.

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 +15,8 @@ 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) {
+        subroutineVariables.add(Pair(decl.name, decl))
+        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()
         }
@@ -40,36 +42,47 @@ internal class BeforeAsmGenerationAstChanger(val program: Program, val errors: E
         return noModifications
     }
 
+    private val subroutineVariables = mutableListOf<Pair<String, VarDecl>>()
+
+    override fun before(subroutine: Subroutine, parent: Node): Iterable<IAstModification> {
+        subroutineVariables.clear()
+        return noModifications
+    }
+
     override fun after(scope: AnonymousScope, parent: Node): Iterable<IAstModification> {
         val decls = scope.statements.filterIsInstance<VarDecl>()
+        subroutineVariables.addAll(decls.map { Pair(it.name, it) })
+
         val sub = scope.definingSubroutine()
         if (sub != null) {
-            val existingVariables = sub.statements.filterIsInstance<VarDecl>().associateBy { it.name }
+            // move vardecls of the scope into the upper scope. Make sure the position remains the same!
-            var conflicts = false
-            decls.forEach {
-                val existing = existingVariables[it.name]
-                if (existing != null) {
-                    errors.err("variable ${it.name} already defined in subroutine ${sub.name} at ${existing.position}", it.position)
-                    conflicts = true
-                }
-            }
-            if (!conflicts) {
             val numericVarsWithValue = decls.filter { it.value != null && it.datatype in NumericDatatypes }
-                return numericVarsWithValue.map {
+            val replaceVardecls =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)
                 val target = AssignTarget(IdentifierReference(listOf(it.name), it.position), null, null, it.position)
                 val assign = Assignment(target, initValue, it.position)
                 initValue.parent = assign
-                    IAstModification.InsertFirst(assign, scope)
+                IAstModification.ReplaceNode(it, assign, scope)
-                } +  decls.map { IAstModification.ReplaceNode(it, NopStatement(it.position), scope) } +
-                     decls.map { IAstModification.InsertFirst(it, sub) }    // move it up to the subroutine
             }
+            val moveVardeclsUp = decls.map { IAstModification.InsertFirst(it, sub) }
+            return replaceVardecls + moveVardeclsUp
         }
         return noModifications
     }
 
     override fun after(subroutine: Subroutine, parent: Node): Iterable<IAstModification> {
+        val firstDeclarations = mutableMapOf<String, VarDecl>()
+        for(decl in subroutineVariables) {
+            val existing = firstDeclarations[decl.first]
+            if(existing!=null && existing !== decl.second) {
+                errors.err("variable ${decl.first} already defined in subroutine ${subroutine.name} at ${existing.position}", decl.second.position)
+            } else {
+                firstDeclarations[decl.first] = decl.second
+            }
+        }
+
+
         // add the implicit return statement at the end (if it's not there yet), but only if it's not a kernel routine.
         // and if an assembly block doesn't contain a rts/rti, and some other situations.
         val mods = mutableListOf<IAstModification>()
@@ -114,6 +127,18 @@ 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/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
@@ -78,7 +91,7 @@ fun compileProgram(filepath: Path,
 }
 
 private fun parseImports(filepath: Path, errors: ErrorReporter): Triple<Program, CompilationOptions, List<Path>> {
-    println("Parsing...")
+    println("Compiler target: ${CompilationTarget.instance.name}. Parsing...")
     val importer = ModuleImporter()
     val programAst = Program(moduleName(filepath.fileName), mutableListOf())
     importer.importModule(programAst, filepath)
@@ -90,8 +103,8 @@ private fun parseImports(filepath: Path, errors: ErrorReporter): Triple<Program,
     if (compilerOptions.launcher == LauncherType.BASIC && compilerOptions.output != OutputType.PRG)
         throw ParsingFailedError("${programAst.modules.first().position} BASIC launcher requires output type PRG.")
 
-    // depending on the mach9ine and compiler options we may have to include some libraries
+    // depending on the machine and compiler options we may have to include some libraries
-    CompilationTarget.machine.importLibs(compilerOptions, importer, programAst)
+    CompilationTarget.instance.machine.importLibs(compilerOptions, importer, programAst)
 
     // always import prog8lib and math
     importer.importLibraryModule(programAst, "math")
@@ -112,7 +125,7 @@ private fun determineCompilationOptions(program: Program): CompilationOptions {
             as? Directive)?.args?.single()?.name?.toUpperCase()
     val allOptions = program.modules.flatMap { it.statements }.filter { it is Directive && it.directive == "%option" }.flatMap { (it as Directive).args }.toSet()
     val floatsEnabled = allOptions.any { it.name == "enable_floats" }
-    val zpType: ZeropageType =
+    var zpType: ZeropageType =
             if (zpoption == null)
                 if(floatsEnabled) ZeropageType.FLOATSAFE else ZeropageType.KERNALSAFE
             else
@@ -122,6 +135,12 @@ private fun determineCompilationOptions(program: Program): CompilationOptions {
                     ZeropageType.KERNALSAFE
                     // error will be printed by the astchecker
                 }
+
+    if (zpType==ZeropageType.FLOATSAFE && CompilationTarget.instance.name == Cx16Target.name) {
+        System.err.println("Warning: Cx16 target must use zp option basicsafe instead of floatsafe")
+        zpType = ZeropageType.BASICSAFE
+    }
+
     val zpReserved = mainModule.statements
             .asSequence()
             .filter { it is Directive && it.directive == "%zpreserved" }
@@ -129,6 +148,15 @@ private fun determineCompilationOptions(program: Program): CompilationOptions {
             .map { it[0].int!!..it[1].int!! }
             .toList()
 
+    if(outputType!=null && !OutputType.values().any {it.name==outputType}) {
+        System.err.println("invalid output type $outputType")
+        exitProcess(1)
+    }
+    if(launcherType!=null && !LauncherType.values().any {it.name==launcherType}) {
+        System.err.println("invalid launcher type $launcherType")
+        exitProcess(1)
+    }
+
     return CompilationOptions(
             if (outputType == null) OutputType.PRG else OutputType.valueOf(outputType),
             if (launcherType == null) LauncherType.BASIC else LauncherType.valueOf(launcherType),
@@ -138,7 +166,7 @@ private fun determineCompilationOptions(program: Program): CompilationOptions {
 
 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)
@@ -160,7 +188,7 @@ private fun optimizeAst(programAst: Program, errors: ErrorReporter) {
         // keep optimizing expressions and statements until no more steps remain
         val optsDone1 = programAst.simplifyExpressions()
         val optsDone2 = programAst.optimizeStatements(errors)
-        programAst.constantFold(errors) // because simplified statements and expressions could give rise to more constants that can be folded away:
+        programAst.constantFold(errors) // because simplified statements and expressions could give rise to more constants that can be folded away
         errors.handle()
         if (optsDone1 + optsDone2 == 0)
             break
@@ -191,11 +219,11 @@ private fun writeAssembly(programAst: Program, errors: ErrorReporter, outputDir:
 
     // printAst(programAst)
 
-    CompilationTarget.machine.initializeZeropage(compilerOptions)
+    CompilationTarget.instance.machine.initializeZeropage(compilerOptions)
-    val assembly = CompilationTarget.asmGenerator(
+    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 initProcName: String?
+    val resetProcName: String?
+
     companion object {
-        lateinit var name: String
+        lateinit var instance: CompilationTarget
-        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
     }
 }
+
+
+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 initProcName = "c64.init_system"
+    override val resetProcName = "c64.reset_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 initProcName = "cx16.init_system"
+    override val resetProcName = "cx16.reset_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)
 
@@ -38,28 +37,28 @@ internal object C64MachineDefinition: IMachineDefinition {
         val mflpt5 = Mflpt5.fromNumber(number)
         val floatbytes = shortArrayOf(mflpt5.b0, mflpt5.b1, mflpt5.b2, mflpt5.b3, mflpt5.b4)
         when {
-            floatbytes.contentEquals(shortArrayOf(0x00, 0x00, 0x00, 0x00, 0x00)) -> return "c64flt.ZERO"        // not a ROM const
+            floatbytes.contentEquals(shortArrayOf(0x00, 0x00, 0x00, 0x00, 0x00)) -> return "floats.ZERO"        // not a ROM const
-            floatbytes.contentEquals(shortArrayOf(0x82, 0x49, 0x0f, 0xda, 0xa1)) -> return "c64flt.FL_PIVAL"
+            floatbytes.contentEquals(shortArrayOf(0x82, 0x49, 0x0f, 0xda, 0xa1)) -> return "floats.FL_PIVAL"
-            floatbytes.contentEquals(shortArrayOf(0x90, 0x80, 0x00, 0x00, 0x00)) -> return "c64flt.FL_N32768"
+            floatbytes.contentEquals(shortArrayOf(0x90, 0x80, 0x00, 0x00, 0x00)) -> return "floats.FL_N32768"
-            floatbytes.contentEquals(shortArrayOf(0x81, 0x00, 0x00, 0x00, 0x00)) -> return "c64flt.FL_FONE"
+            floatbytes.contentEquals(shortArrayOf(0x81, 0x00, 0x00, 0x00, 0x00)) -> return "floats.FL_FONE"
-            floatbytes.contentEquals(shortArrayOf(0x80, 0x35, 0x04, 0xf3, 0x34)) -> return "c64flt.FL_SQRHLF"
+            floatbytes.contentEquals(shortArrayOf(0x80, 0x35, 0x04, 0xf3, 0x34)) -> return "floats.FL_SQRHLF"
-            floatbytes.contentEquals(shortArrayOf(0x81, 0x35, 0x04, 0xf3, 0x34)) -> return "c64flt.FL_SQRTWO"
+            floatbytes.contentEquals(shortArrayOf(0x81, 0x35, 0x04, 0xf3, 0x34)) -> return "floats.FL_SQRTWO"
-            floatbytes.contentEquals(shortArrayOf(0x80, 0x80, 0x00, 0x00, 0x00)) -> return "c64flt.FL_NEGHLF"
+            floatbytes.contentEquals(shortArrayOf(0x80, 0x80, 0x00, 0x00, 0x00)) -> return "floats.FL_NEGHLF"
-            floatbytes.contentEquals(shortArrayOf(0x80, 0x31, 0x72, 0x17, 0xf8)) -> return "c64flt.FL_LOG2"
+            floatbytes.contentEquals(shortArrayOf(0x80, 0x31, 0x72, 0x17, 0xf8)) -> return "floats.FL_LOG2"
-            floatbytes.contentEquals(shortArrayOf(0x84, 0x20, 0x00, 0x00, 0x00)) -> return "c64flt.FL_TENC"
+            floatbytes.contentEquals(shortArrayOf(0x84, 0x20, 0x00, 0x00, 0x00)) -> return "floats.FL_TENC"
-            floatbytes.contentEquals(shortArrayOf(0x9e, 0x6e, 0x6b, 0x28, 0x00)) -> return "c64flt.FL_NZMIL"
+            floatbytes.contentEquals(shortArrayOf(0x9e, 0x6e, 0x6b, 0x28, 0x00)) -> return "floats.FL_NZMIL"
-            floatbytes.contentEquals(shortArrayOf(0x80, 0x00, 0x00, 0x00, 0x00)) -> return "c64flt.FL_FHALF"
+            floatbytes.contentEquals(shortArrayOf(0x80, 0x00, 0x00, 0x00, 0x00)) -> return "floats.FL_FHALF"
-            floatbytes.contentEquals(shortArrayOf(0x81, 0x38, 0xaa, 0x3b, 0x29)) -> return "c64flt.FL_LOGEB2"
+            floatbytes.contentEquals(shortArrayOf(0x81, 0x38, 0xaa, 0x3b, 0x29)) -> return "floats.FL_LOGEB2"
-            floatbytes.contentEquals(shortArrayOf(0x81, 0x49, 0x0f, 0xda, 0xa2)) -> return "c64flt.FL_PIHALF"
+            floatbytes.contentEquals(shortArrayOf(0x81, 0x49, 0x0f, 0xda, 0xa2)) -> return "floats.FL_PIHALF"
-            floatbytes.contentEquals(shortArrayOf(0x83, 0x49, 0x0f, 0xda, 0xa2)) -> return "c64flt.FL_TWOPI"
+            floatbytes.contentEquals(shortArrayOf(0x83, 0x49, 0x0f, 0xda, 0xa2)) -> return "floats.FL_TWOPI"
-            floatbytes.contentEquals(shortArrayOf(0x7f, 0x00, 0x00, 0x00, 0x00)) -> return "c64flt.FL_FR4"
+            floatbytes.contentEquals(shortArrayOf(0x7f, 0x00, 0x00, 0x00, 0x00)) -> return "floats.FL_FR4"
             else -> {
                 // attempt to correct for a few rounding issues
                 when (number.toBigDecimal().setScale(10, RoundingMode.HALF_DOWN).toDouble()) {
-                    3.1415926536 -> return "c64flt.FL_PIVAL"
+                    3.1415926536 -> return "floats.FL_PIVAL"
-                    1.4142135624 -> return "c64flt.FL_SQRTWO"
+                    1.4142135624 -> return "floats.FL_SQRTWO"
-                    0.7071067812 -> return "c64flt.FL_SQRHLF"
+                    0.7071067812 -> return "floats.FL_SQRHLF"
-                    0.6931471806 -> return "c64flt.FL_LOG2"
+                    0.6931471806 -> return "floats.FL_LOG2"
                     else -> {}
                 }
             }
@@ -69,7 +68,7 @@ internal object C64MachineDefinition: IMachineDefinition {
 
     override fun importLibs(compilerOptions: CompilationOptions, importer: ModuleImporter, program: Program) {
         if (compilerOptions.launcher == LauncherType.BASIC || compilerOptions.output == OutputType.PRG)
-            importer.importLibraryModule(program, "c64lib")
+            importer.importLibraryModule(program, "syslib")
     }
 
     override fun launchEmulator(programName: String) {
@@ -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) {
@@ -149,7 +147,7 @@ internal object C64MachineDefinition: IMachineDefinition {
                 if (options.zeropage == ZeropageType.FLOATSAFE) {
                     // remove the zero page locations used for floating point operations from the free list
                     free.removeAll(listOf(
-                            0x12, 0x26, 0x27, 0x28, 0x29, 0x2a,
+                            0x10, 0x11, 0x12, 0x26, 0x27, 0x28, 0x29, 0x2a,
                             0x57, 0x58, 0x59, 0x5a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, 0x60,
                             0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68,
                             0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, 0x70, 0x71, 0x72,
@@ -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 zp = CompilationTarget.instance.machine.zeropage
-        val initproc = CompilationTarget.machine.initSystemProcname
         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_LO = ${CompilationTarget.instance.machine.ESTACK_LO.toHex()}")
-        out("P8ESTACK_HI = ${CompilationTarget.machine.ESTACK_HI.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())
+                if(!CompilationTarget.instance.initProcName.isNullOrEmpty())
-                    out("  jsr  $initproc")
+                    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())
+                if(!CompilationTarget.instance.initProcName.isNullOrEmpty())
-                    out("  jsr  $initproc")
+                    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("  jmp  ${CompilationTarget.instance.resetProcName}")
             }
         }
     }
@@ -170,7 +172,7 @@ internal class AsmGen(private val program: Program,
         // the global list of all floating point constants for the whole program
         out("; global float constants")
         for (flt in globalFloatConsts) {
-            val floatFill = CompilationTarget.machine.getFloat(flt.key).makeFloatFillAsm()
+            val floatFill = CompilationTarget.instance.machine.getFloat(flt.key).makeFloatFillAsm()
             val floatvalue = flt.key
             out("${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))
@@ -353,7 +355,10 @@ internal class AsmGen(private val program: Program,
         out("\n; memdefs and kernel subroutines")
         val memvars = statements.filterIsInstance<VarDecl>().filter { it.type==VarDeclType.MEMORY || it.type==VarDeclType.CONST }
         for(m in memvars) {
+            if(m.value is NumericLiteralValue)
                 out("  ${m.name} = ${(m.value as NumericLiteralValue).number.toHex()}")
+            else
+                out("  ${m.name} = ${asmVariableName((m.value as AddressOf).identifier)}")
         }
         val asmSubs = statements.filterIsInstance<Subroutine>().filter { it.isAsmSubroutine }
         for(sub in asmSubs) {
@@ -474,7 +479,7 @@ 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 +553,35 @@ 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)
+                if (CompilationTarget.instance.machine.cpu == CpuType.CPU65c02) out("  phx")
-                    out("  phx")
+                else {
-                else
+                    val save = makeLabel("saveX")
-                    out("  stx  P8ZP_SCRATCH_REG_X")
+                    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")
         }
     }
 
@@ -565,14 +589,22 @@ internal class AsmGen(private val program: Program,
         when(register) {
             CpuRegister.A -> out("  pla")
             CpuRegister.X -> {
-                if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
+                if (CompilationTarget.instance.machine.cpu == CpuType.CPU65c02) out("  plx")
-                    out("  plx")
+                else {
-                else
+                    val save = saveRegisterLabels.pop()
-                    out("  ldx  P8ZP_SCRATCH_REG_X")
+                    out("  ldx  $save")
-            }
-            CpuRegister.Y -> out("  pla | tay")
                 }
             }
+            CpuRegister.Y -> {
+                if (CompilationTarget.instance.machine.cpu == CpuType.CPU65c02) out("  ply")
+                else {
+                    val save = saveRegisterLabels.pop()
+                    out("  ldy  $save")
+                }
+            }
+        }
+    }
+
 
     internal fun translate(stmt: Statement) {
         outputSourceLine(stmt)
@@ -684,7 +716,7 @@ 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 +754,7 @@ 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")
                 }
             }
@@ -792,6 +824,26 @@ internal class AsmGen(private val program: Program,
             }
 
     private fun translate(stmt: IfStatement) {
+        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")
@@ -803,6 +855,8 @@ internal class AsmGen(private val program: Program,
                 translate(stmt.elsepart)
                 out(endLabel)
             }
+        }
+    }
 
     private fun translateTestStack(dataType: DataType) {
         when(dataType) {

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

@@ -152,7 +152,7 @@ private fun optimizeSameAssignments(linesByFourteen: List<List<IndexedValue<Stri
 
         if(first.startsWith("lda") && second.startsWith("ldy") && third.startsWith("sta") && fourth.startsWith("sty") &&
                 fifth.startsWith("lda") && sixth.startsWith("ldy") &&
-                (seventh.startsWith("jsr  c64flt.copy_float") || seventh.startsWith("jsr  cx16flt.copy_float"))) {
+                (seventh.startsWith("jsr  floats.copy_float") || seventh.startsWith("jsr  cx16flt.copy_float"))) {
 
             val nineth = pair[8].value.trimStart()
             val tenth = pair[9].value.trimStart()
@@ -163,7 +163,7 @@ private fun optimizeSameAssignments(linesByFourteen: List<List<IndexedValue<Stri
 
             if(eighth.startsWith("lda") && nineth.startsWith("ldy") && tenth.startsWith("sta") && eleventh.startsWith("sty") &&
                     twelveth.startsWith("lda") && thirteenth.startsWith("ldy") &&
-                    (fourteenth.startsWith("jsr  c64flt.copy_float") || fourteenth.startsWith("jsr  cx16flt.copy_float"))) {
+                    (fourteenth.startsWith("jsr  floats.copy_float") || fourteenth.startsWith("jsr  cx16flt.copy_float"))) {
 
                 if(first.substring(4) == eighth.substring(4) && second.substring(4)==nineth.substring(4)) {
                     // identical float init

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

@@ -4,8 +4,14 @@ import prog8.ast.IFunctionCall
 import prog8.ast.Program
 import prog8.ast.base.*
 import prog8.ast.expressions.*
+import prog8.ast.statements.DirectMemoryWrite
 import prog8.ast.statements.FunctionCallStatement
 import prog8.compiler.AssemblyError
+import prog8.compiler.target.c64.codegen.assignment.AsmAssignSource
+import prog8.compiler.target.c64.codegen.assignment.AsmAssignTarget
+import prog8.compiler.target.c64.codegen.assignment.AsmAssignment
+import prog8.compiler.target.c64.codegen.assignment.SourceStorageKind
+import prog8.compiler.target.c64.codegen.assignment.TargetStorageKind
 import prog8.compiler.toHex
 import prog8.functions.FSignature
 
@@ -342,7 +348,7 @@ internal class BuiltinFunctionsAsmGen(private val program: Program, private val
 
     private fun funcVariousFloatFuncs(fcall: IFunctionCall, func: FSignature, functionName: String) {
         translateFunctionArguments(fcall.args, func)
-        asmgen.out("  jsr  c64flt.func_$functionName")
+        asmgen.out("  jsr  floats.func_$functionName")
     }
 
     private fun funcSgn(fcall: IFunctionCall, func: FSignature) {
@@ -353,7 +359,7 @@ internal class BuiltinFunctionsAsmGen(private val program: Program, private val
             DataType.BYTE -> asmgen.out("  jsr  math.sign_b")
             DataType.UWORD -> asmgen.out("  jsr  math.sign_uw")
             DataType.WORD -> asmgen.out("  jsr  math.sign_w")
-            DataType.FLOAT -> asmgen.out("  jsr  c64flt.sign_f")
+            DataType.FLOAT -> asmgen.out("  jsr  floats.sign_f")
             else -> throw AssemblyError("weird type $dt")
         }
     }
@@ -364,7 +370,7 @@ internal class BuiltinFunctionsAsmGen(private val program: Program, private val
         when (dt.typeOrElse(DataType.STRUCT)) {
             DataType.ARRAY_B, DataType.ARRAY_UB, DataType.STR -> asmgen.out("  jsr  prog8_lib.func_${functionName}_b")
             DataType.ARRAY_UW, DataType.ARRAY_W -> asmgen.out("  jsr  prog8_lib.func_${functionName}_w")
-            DataType.ARRAY_F -> asmgen.out("  jsr  c64flt.func_${functionName}_f")
+            DataType.ARRAY_F -> asmgen.out("  jsr  floats.func_${functionName}_f")
             else -> throw AssemblyError("weird type $dt")
         }
     }
@@ -377,7 +383,7 @@ internal class BuiltinFunctionsAsmGen(private val program: Program, private val
             DataType.ARRAY_B -> asmgen.out("  jsr  prog8_lib.func_${functionName}_b")
             DataType.ARRAY_UW -> asmgen.out("  jsr  prog8_lib.func_${functionName}_uw")
             DataType.ARRAY_W -> asmgen.out("  jsr  prog8_lib.func_${functionName}_w")
-            DataType.ARRAY_F -> asmgen.out("  jsr  c64flt.func_${functionName}_f")
+            DataType.ARRAY_F -> asmgen.out("  jsr  floats.func_${functionName}_f")
             else -> throw AssemblyError("weird type $dt")
         }
     }
@@ -395,12 +401,14 @@ internal class BuiltinFunctionsAsmGen(private val program: Program, private val
     private fun funcSwap(fcall: IFunctionCall) {
         val first = fcall.args[0]
         val second = fcall.args[1]
+
+        // optimized simple case: swap two variables
         if(first is IdentifierReference && second is IdentifierReference) {
             val firstName = asmgen.asmVariableName(first)
             val secondName = asmgen.asmVariableName(second)
             val dt = first.inferType(program)
             if(dt.istype(DataType.BYTE) || dt.istype(DataType.UBYTE)) {
-                asmgen.out(" ldy  $firstName |  lda  $secondName |  sta  $firstName |  tya |  sta  $secondName")
+                asmgen.out(" ldy  $firstName |  lda  $secondName |  sta  $firstName |  sty  $secondName")
                 return
             }
             if(dt.istype(DataType.WORD) || dt.istype(DataType.UWORD)) {
@@ -426,14 +434,196 @@ internal class BuiltinFunctionsAsmGen(private val program: Program, private val
                     sta  P8ZP_SCRATCH_W2
                     lda  #>$secondName
                     sta  P8ZP_SCRATCH_W2+1
-                    jsr  c64flt.swap_floats
+                    jsr  floats.swap_floats
                 """)
                 return
             }
         }
 
-        // other types of swap() calls should have been replaced by a different statement sequence involving a temp variable
+        // optimized simple case: swap two memory locations
-        throw AssemblyError("no asm generation for swap funccall $fcall")
+        if(first is DirectMemoryRead && second is DirectMemoryRead) {
+            val addr1 = (first.addressExpression as? NumericLiteralValue)?.number?.toHex()
+            val addr2 = (second.addressExpression as? NumericLiteralValue)?.number?.toHex()
+            val name1 = if(first.addressExpression is IdentifierReference) asmgen.asmVariableName(first.addressExpression as IdentifierReference) else null
+            val name2 = if(second.addressExpression is IdentifierReference) asmgen.asmVariableName(second.addressExpression as IdentifierReference) else null
+
+            when {
+                addr1!=null && addr2!=null -> {
+                    asmgen.out("  ldy  $addr1 |  lda  $addr2 |  sta  $addr1 |  sty  $addr2")
+                    return
+                }
+                addr1!=null && name2!=null -> {
+                    asmgen.out("  ldy  $addr1 |  lda  $name2 |  sta  $addr1 |  sty  $name2")
+                    return
+                }
+                name1!=null && addr2 != null -> {
+                    asmgen.out("  ldy  $name1 |  lda  $addr2 |  sta  $name1 |  sty  $addr2")
+                    return
+                }
+                name1!=null && name2!=null -> {
+                    asmgen.out("  ldy  $name1 |  lda  $name2 |  sta  $name1 |  sty  $name2")
+                    return
+                }
+            }
+        }
+
+        if(first is ArrayIndexedExpression && second is ArrayIndexedExpression) {
+            val indexValue1 = first.arrayspec.index as? NumericLiteralValue
+            val indexName1 = first.arrayspec.index as? IdentifierReference
+            val indexValue2 = second.arrayspec.index as? NumericLiteralValue
+            val indexName2 = second.arrayspec.index as? IdentifierReference
+            val arrayVarName1 = asmgen.asmVariableName(first.identifier)
+            val arrayVarName2 = asmgen.asmVariableName(second.identifier)
+            val elementDt = first.inferType(program).typeOrElse(DataType.STRUCT)
+
+            if(indexValue1!=null && indexValue2!=null) {
+                swapArrayValues(elementDt, arrayVarName1, indexValue1, arrayVarName2, indexValue2)
+                return
+            } else if(indexName1!=null && indexName2!=null) {
+                swapArrayValues(elementDt, arrayVarName1, indexName1, arrayVarName2, indexName2)
+                return
+            }
+        }
+
+        // all other types of swap() calls are done via the evaluation stack
+        fun targetFromExpr(expr: Expression, datatype: DataType): AsmAssignTarget {
+            return when (expr) {
+                is IdentifierReference -> AsmAssignTarget(TargetStorageKind.VARIABLE, program, asmgen, datatype, variable=expr)
+                is ArrayIndexedExpression -> AsmAssignTarget(TargetStorageKind.ARRAY, program, asmgen, datatype, array = expr)
+                is DirectMemoryRead -> AsmAssignTarget(TargetStorageKind.MEMORY, program, asmgen, datatype, memory = DirectMemoryWrite(expr.addressExpression, expr.position))
+                else -> throw AssemblyError("invalid expression object $expr")
+            }
+        }
+
+        asmgen.translateExpression(first)
+        asmgen.translateExpression(second)
+        val datatype = first.inferType(program).typeOrElse(DataType.STRUCT)
+        val assignFirst = AsmAssignment(
+                AsmAssignSource(SourceStorageKind.STACK, program, datatype),
+                targetFromExpr(first, datatype),
+                false, first.position
+        )
+        val assignSecond = AsmAssignment(
+                AsmAssignSource(SourceStorageKind.STACK, program, datatype),
+                targetFromExpr(second, datatype),
+                false, second.position
+        )
+        asmgen.translateNormalAssignment(assignFirst)
+        asmgen.translateNormalAssignment(assignSecond)
+    }
+
+    private fun swapArrayValues(elementDt: DataType, arrayVarName1: String, indexValue1: NumericLiteralValue, arrayVarName2: String, indexValue2: NumericLiteralValue) {
+        val index1 = indexValue1.number.toInt() * elementDt.memorySize()
+        val index2 = indexValue2.number.toInt() * elementDt.memorySize()
+        when(elementDt) {
+            DataType.UBYTE, DataType.BYTE -> {
+                asmgen.out("""
+                    lda  $arrayVarName1+$index1
+                    ldy  $arrayVarName2+$index2
+                    sta  $arrayVarName2+$index2
+                    sty  $arrayVarName1+$index1
+                """)
+            }
+            DataType.UWORD, DataType.WORD -> {
+                asmgen.out("""
+                    lda  $arrayVarName1+$index1
+                    ldy  $arrayVarName2+$index2
+                    sta  $arrayVarName2+$index2
+                    sty  $arrayVarName1+$index1
+                    lda  $arrayVarName1+$index1+1
+                    ldy  $arrayVarName2+$index2+1
+                    sta  $arrayVarName2+$index2+1
+                    sty  $arrayVarName1+$index1+1
+                """)
+            }
+            DataType.FLOAT -> {
+                asmgen.out("""
+                    lda  #<(${arrayVarName1}+$index1)
+                    sta  P8ZP_SCRATCH_W1
+                    lda  #>(${arrayVarName1}+$index1)
+                    sta  P8ZP_SCRATCH_W1+1
+                    lda  #<(${arrayVarName2}+$index2)
+                    sta  P8ZP_SCRATCH_W2
+                    lda  #>(${arrayVarName2}+$index2)
+                    sta  P8ZP_SCRATCH_W2+1
+                    jsr  floats.swap_floats
+                """)
+            }
+            else -> throw AssemblyError("invalid aray elt type")
+        }
+    }
+
+    private fun swapArrayValues(elementDt: DataType, arrayVarName1: String, indexName1: IdentifierReference, arrayVarName2: String, indexName2: IdentifierReference) {
+        val idxAsmName1 = asmgen.asmVariableName(indexName1)
+        val idxAsmName2 = asmgen.asmVariableName(indexName2)
+        when(elementDt) {
+            DataType.UBYTE, DataType.BYTE -> {
+                asmgen.out("""
+                    stx  P8ZP_SCRATCH_REG
+                    ldx  $idxAsmName1
+                    ldy  $idxAsmName2
+                    lda  $arrayVarName1,x
+                    pha
+                    lda  $arrayVarName2,y
+                    sta  $arrayVarName1,x
+                    pla
+                    sta  $arrayVarName2,y
+                    ldx  P8ZP_SCRATCH_REG
+                """)
+            }
+            DataType.UWORD, DataType.WORD -> {
+                asmgen.out("""
+                    stx  P8ZP_SCRATCH_REG
+                    lda  $idxAsmName1
+                    asl  a
+                    tax
+                    lda  $idxAsmName2
+                    asl  a
+                    tay
+                    lda  $arrayVarName1,x
+                    pha
+                    lda  $arrayVarName2,y
+                    sta  $arrayVarName1,x
+                    pla
+                    sta  $arrayVarName2,y
+                    lda  $arrayVarName1+1,x
+                    pha
+                    lda  $arrayVarName2+1,y
+                    sta  $arrayVarName1+1,x
+                    pla
+                    sta  $arrayVarName2+1,y                    
+                    ldx  P8ZP_SCRATCH_REG
+                """)
+            }
+            DataType.FLOAT -> {
+                asmgen.out("""
+                    lda  #>$arrayVarName1
+                    sta  P8ZP_SCRATCH_W1+1
+                    lda  $idxAsmName1
+                    asl  a
+                    asl  a
+                    clc
+                    adc  $idxAsmName1
+                    adc  #<$arrayVarName1
+                    sta  P8ZP_SCRATCH_W1
+                    bcc  +
+                    inc  P8ZP_SCRATCH_W1+1
++                   lda  #>$arrayVarName2
+                    sta  P8ZP_SCRATCH_W2+1
+                    lda  $idxAsmName2
+                    asl  a
+                    asl  a
+                    clc
+                    adc  $idxAsmName2
+                    adc  #<$arrayVarName2
+                    sta  P8ZP_SCRATCH_W2
+                    bcc  +
+                    inc  P8ZP_SCRATCH_W2+1
++                   jsr  floats.swap_floats                                   
+                """)
+            }
+            else -> throw AssemblyError("invalid aray elt type")
+        }
     }
 
     private fun funcAbs(fcall: IFunctionCall, func: FSignature) {
@@ -442,7 +632,7 @@ internal class BuiltinFunctionsAsmGen(private val program: Program, private val
         when (dt.typeOrElse(DataType.STRUCT)) {
             in ByteDatatypes -> asmgen.out("  jsr  prog8_lib.abs_b")
             in WordDatatypes -> asmgen.out("  jsr  prog8_lib.abs_w")
-            DataType.FLOAT -> asmgen.out("  jsr  c64flt.abs_f")
+            DataType.FLOAT -> asmgen.out("  jsr  floats.abs_f")
             else -> throw AssemblyError("weird type")
         }
     }

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")
@@ -85,12 +85,12 @@ 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")
                     }
-                    DataType.FLOAT -> asmgen.out(" jsr  c64flt.stack_ub2float")
+                    DataType.FLOAT -> asmgen.out(" jsr  floats.stack_ub2float")
                     in PassByReferenceDatatypes -> throw AssemblyError("cannot cast to a pass-by-reference datatype")
                     else -> throw AssemblyError("weird type")
                 }
@@ -103,16 +103,11 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                         asmgen.out(""" 
                             lda  P8ESTACK_LO+1,x
                             ora  #$7f
-                            bmi  +""")
+                            bmi  +
-                        if(CompilationTarget.machine.cpu==CpuType.CPU65c02)
-                            asmgen.out("""
-+                               stz  P8ESTACK_HI+1,x""")
-                        else
-                            asmgen.out("""
                             lda  #0
 +                           sta  P8ESTACK_HI+1,x""")
                     }
-                    DataType.FLOAT -> asmgen.out(" jsr  c64flt.stack_b2float")
+                    DataType.FLOAT -> asmgen.out(" jsr  floats.stack_b2float")
                     in PassByReferenceDatatypes -> throw AssemblyError("cannot cast to a pass-by-reference datatype")
                     else -> throw AssemblyError("weird type")
                 }
@@ -121,7 +116,7 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                 when(expr.type) {
                     DataType.BYTE, DataType.UBYTE -> {}
                     DataType.WORD, DataType.UWORD -> {}
-                    DataType.FLOAT -> asmgen.out(" jsr  c64flt.stack_uw2float")
+                    DataType.FLOAT -> asmgen.out(" jsr  floats.stack_uw2float")
                     in PassByReferenceDatatypes -> throw AssemblyError("cannot cast to a pass-by-reference datatype")
                     else -> throw AssemblyError("weird type")
                 }
@@ -130,17 +125,17 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                 when(expr.type) {
                     DataType.BYTE, DataType.UBYTE -> {}
                     DataType.WORD, DataType.UWORD -> {}
-                    DataType.FLOAT -> asmgen.out(" jsr  c64flt.stack_w2float")
+                    DataType.FLOAT -> asmgen.out(" jsr  floats.stack_w2float")
                     in PassByReferenceDatatypes -> throw AssemblyError("cannot cast to a pass-by-reference datatype")
                     else -> throw AssemblyError("weird type")
                 }
             }
             DataType.FLOAT -> {
                 when(expr.type) {
-                    DataType.UBYTE -> asmgen.out(" jsr  c64flt.stack_float2uw")
+                    DataType.UBYTE -> asmgen.out(" jsr  floats.stack_float2uw")
-                    DataType.BYTE -> asmgen.out(" jsr  c64flt.stack_float2w")
+                    DataType.BYTE -> asmgen.out(" jsr  floats.stack_float2w")
-                    DataType.UWORD -> asmgen.out(" jsr  c64flt.stack_float2uw")
+                    DataType.UWORD -> asmgen.out(" jsr  floats.stack_float2uw")
-                    DataType.WORD -> asmgen.out(" jsr  c64flt.stack_float2w")
+                    DataType.WORD -> asmgen.out(" jsr  floats.stack_float2w")
                     DataType.FLOAT -> {}
                     in PassByReferenceDatatypes -> throw AssemblyError("cannot cast to a pass-by-reference datatype")
                     else -> throw AssemblyError("weird type")
@@ -187,7 +182,7 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
             """)
             DataType.FLOAT -> {
                 val floatConst = asmgen.getFloatAsmConst(expr.number.toDouble())
-                asmgen.out(" lda  #<$floatConst |  ldy  #>$floatConst |  jsr  c64flt.push_float")
+                asmgen.out(" lda  #<$floatConst |  ldy  #>$floatConst |  jsr  floats.push_float")
             }
             else -> throw AssemblyError("weird type")
         }
@@ -203,7 +198,7 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                 asmgen.out("  lda  $varname  |  sta  P8ESTACK_LO,x  |  lda  $varname+1 |  sta  P8ESTACK_HI,x |  dex")
             }
             DataType.FLOAT -> {
-                asmgen.out(" lda  #<$varname |  ldy  #>$varname|  jsr  c64flt.push_float")
+                asmgen.out(" lda  #<$varname |  ldy  #>$varname|  jsr  floats.push_float")
             }
             in IterableDatatypes -> {
                 asmgen.out("  lda  #<$varname  |  sta  P8ESTACK_LO,x  |  lda  #>$varname |  sta  P8ESTACK_HI,x |  dex")
@@ -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
                                 }
                             }
@@ -376,7 +368,7 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                 when(type) {
                     in ByteDatatypes -> asmgen.out("  jsr  prog8_lib.neg_b")
                     in WordDatatypes -> asmgen.out("  jsr  prog8_lib.neg_w")
-                    DataType.FLOAT -> asmgen.out("  jsr  c64flt.neg_f")
+                    DataType.FLOAT -> asmgen.out("  jsr  floats.neg_f")
                     else -> throw AssemblyError("weird type")
                 }
             }
@@ -417,7 +409,7 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                     asmgen.out("  lda  $arrayVarName+$indexValue |  sta  P8ESTACK_LO,x |  lda  $arrayVarName+$indexValue+1 |  sta  P8ESTACK_HI,x |  dex")
                 }
                 DataType.FLOAT -> {
-                    asmgen.out("  lda  #<$arrayVarName+$indexValue |  ldy  #>$arrayVarName+$indexValue |  jsr  c64flt.push_float")
+                    asmgen.out("  lda  #<$arrayVarName+$indexValue |  ldy  #>$arrayVarName+$indexValue |  jsr  floats.push_float")
                 }
                 else -> throw AssemblyError("weird element type")
             }
@@ -439,7 +431,7 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                         adc  #<$arrayVarName
                         bcc  +
                         iny
-+                       jsr  c64flt.push_float""")
++                       jsr  floats.push_float""")
                 }
                 else -> throw AssemblyError("weird dt")
             }
@@ -521,17 +513,17 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
 
     private fun translateBinaryOperatorFloats(operator: String) {
         when(operator) {
-            "**" -> asmgen.out(" jsr  c64flt.pow_f")
+            "**" -> asmgen.out(" jsr  floats.pow_f")
-            "*" -> asmgen.out("  jsr  c64flt.mul_f")
+            "*" -> asmgen.out("  jsr  floats.mul_f")
-            "/" -> asmgen.out("  jsr  c64flt.div_f")
+            "/" -> asmgen.out("  jsr  floats.div_f")
-            "+" -> asmgen.out("  jsr  c64flt.add_f")
+            "+" -> asmgen.out("  jsr  floats.add_f")
-            "-" -> asmgen.out("  jsr  c64flt.sub_f")
+            "-" -> asmgen.out("  jsr  floats.sub_f")
-            "<" -> asmgen.out("  jsr  c64flt.less_f")
+            "<" -> asmgen.out("  jsr  floats.less_f")
-            ">" -> asmgen.out("  jsr  c64flt.greater_f")
+            ">" -> asmgen.out("  jsr  floats.greater_f")
-            "<=" -> asmgen.out("  jsr  c64flt.lesseq_f")
+            "<=" -> asmgen.out("  jsr  floats.lesseq_f")
-            ">=" -> asmgen.out("  jsr  c64flt.greatereq_f")
+            ">=" -> asmgen.out("  jsr  floats.greatereq_f")
-            "==" -> asmgen.out("  jsr  c64flt.equal_f")
+            "==" -> asmgen.out("  jsr  floats.equal_f")
-            "!=" -> asmgen.out("  jsr  c64flt.notequal_f")
+            "!=" -> asmgen.out("  jsr  floats.notequal_f")
             "%", "<<", ">>", "&", "^", "|", "and", "or", "xor" -> throw AssemblyError("requires integer datatype")
             else -> throw AssemblyError("invalid operator $operator")
         }

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 -> {
+                argi.value.second.stack -> TODO("asmsub @stack parameter")
+                argi.value.second.statusflag == Statusflag.Pc -> {
+                    require(argForCarry == null)
+                    argForCarry = argi
+                }
+                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
+                }
+                argi.value.second.registerOrPair in setOf(RegisterOrPair.A, RegisterOrPair.AY) -> {
+                    require(argForAregister == null)
+                    argForAregister = argi
+                }
+                argi.value.second.registerOrPair == RegisterOrPair.Y -> {
+                    asmgen.out("  ldy  P8ESTACK_LO+${argi.index},x")
+                }
+                else -> throw AssemblyError("weird argument")
+            }
+        }
+
+        if(argForCarry!=null) {
             asmgen.out("""
-                        inx
+                lda  P8ESTACK_LO+${argForCarry.index},x
-                        pha
-                        lda  P8ESTACK_LO,x
                 beq  +
                 sec
                 bcs  ++
 +               clc
-+                       pla""")
++               php""")             // push the status flags
         }
-                regparam.statusflag!=null -> {
+
-                    throw AssemblyError("can only use Carry as status flag parameter")
+        if(argForAregister!=null) {
-                }
+            when(argForAregister.value.second.registerOrPair) {
-                regparam.registerOrPair!=null -> {
+                RegisterOrPair.A -> asmgen.out("  lda  P8ESTACK_LO+${argForAregister.index},x")
-                    val tgt = AsmAssignTarget.fromRegisters(regparam.registerOrPair, program, asmgen)
+                RegisterOrPair.AY -> asmgen.out("  lda  P8ESTACK_LO+${argForAregister.index},x |  ldy  P8ESTACK_HI+${argForAregister.index},x")
-                    val source = AsmAssignSource(SourceStorageKind.STACK, program, tgt.datatype)
+                else -> throw AssemblyError("weird arg")
-                    val asgn = AsmAssignment(source, tgt, false, Position.DUMMY)
-                    asmgen.translateNormalAssignment(asgn)
-                }
-                else -> {}
             }
         }
+
+        if(argForXregister!=null) {
+
+            if(argForAregister!=null)
+                asmgen.out("  pha")
+            when(argForXregister.value.second.registerOrPair) {
+                RegisterOrPair.X -> asmgen.out("  lda  P8ESTACK_LO+${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/PostIncrDecrAsmGen.kt

@@ -33,7 +33,7 @@ internal class PostIncrDecrAsmGen(private val program: Program, private val asmg
                     }
                     DataType.FLOAT -> {
                         asmgen.out("  lda  #<$what |  ldy  #>$what")
-                        asmgen.out(if(incr) "  jsr  c64flt.inc_var_f" else "  jsr  c64flt.dec_var_f")
+                        asmgen.out(if(incr) "  jsr  floats.inc_var_f" else "  jsr  floats.dec_var_f")
                     }
                     else -> throw AssemblyError("need numeric type")
                 }
@@ -90,7 +90,7 @@ internal class PostIncrDecrAsmGen(private val program: Program, private val asmg
                             }
                             DataType.FLOAT -> {
                                 asmgen.out("  lda  #<$asmArrayvarname+$indexValue |  ldy  #>$asmArrayvarname+$indexValue")
-                                asmgen.out(if(incr) "  jsr  c64flt.inc_var_f" else "  jsr  c64flt.dec_var_f")
+                                asmgen.out(if(incr) "  jsr  floats.inc_var_f" else "  jsr  floats.dec_var_f")
                             }
                             else -> throw AssemblyError("need numeric type")
                         }
@@ -121,7 +121,7 @@ internal class PostIncrDecrAsmGen(private val program: Program, private val asmg
                         adc  #<$asmArrayvarname
                         bcc  +
                         iny
-+                       jsr  c64flt.inc_var_f""")
++                       jsr  floats.inc_var_f""")
                             }
                             else -> throw AssemblyError("weird array elt dt")
                         }

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

@@ -65,7 +65,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                         in WordDatatypes ->
                             asmgen.out("  lda  $arrayVarName+$indexValue |  sta  P8ESTACK_LO,x |  lda  $arrayVarName+$indexValue+1 |  sta  P8ESTACK_HI,x | dex")
                         DataType.FLOAT ->
-                            asmgen.out("  lda  #<$arrayVarName+$indexValue |  ldy  #>$arrayVarName+$indexValue |  jsr  c64flt.push_float")
+                            asmgen.out("  lda  #<$arrayVarName+$indexValue |  ldy  #>$arrayVarName+$indexValue |  jsr  floats.push_float")
                         else ->
                             throw AssemblyError("weird array type")
                     }
@@ -87,7 +87,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                                 adc  #<$arrayVarName
                                 bcc  +
                                 iny
-+                               jsr  c64flt.push_float""")
++                               jsr  floats.push_float""")
                         }
                         else ->
                             throw AssemblyError("weird array elt type")
@@ -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 -> {
+                    is TypecastExpression -> assignTypeCastedValue(assign.target, value.type, value.expression, assign)
-//                        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 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 ${value.position}")
+            }
+        }
+
+        // give up, do it via eval stack
+        asmgen.translateExpression(origAssign.source.expression!!)
+        assignStackValue(target)
+    }
+
     private fun assignStackValue(target: AsmAssignTarget) {
         when(target.kind) {
             TargetStorageKind.VARIABLE -> {
@@ -181,7 +217,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                         asmgen.out("""
                             lda  #<${target.asmVarname}
                             ldy  #>${target.asmVarname}
-                            jsr  c64flt.pop_float
+                            jsr  floats.pop_float
                         """)
                     }
                     else -> throw AssemblyError("weird target variable type ${target.datatype}")
@@ -213,7 +249,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                                 asmgen.out("""
                                     lda  #<${target.asmVarname}+$scaledIdx
                                     ldy  #>${target.asmVarname}+$scaledIdx
-                                    jsr  c64flt.pop_float
+                                    jsr  floats.pop_float
                                 """)
                             }
                             else -> throw AssemblyError("weird target variable type ${target.datatype}")
@@ -243,7 +279,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                                     adc  #<${target.asmVarname}
                                     bcc  +
                                     iny
-+                                   jsr  c64flt.pop_float""")
++                                   jsr  floats.pop_float""")
                             }
                             else -> throw AssemblyError("weird dt")
                         }
@@ -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? ${target.origAstTarget?.position}")
                             else -> throw AssemblyError("can't assign word to single byte register")
                         }
                     }
@@ -368,7 +406,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                                     sty  P8ZP_SCRATCH_W1+1
                                     lda  #<${target.asmVarname}+$scaledIdx
                                     ldy  #>${target.asmVarname}+$scaledIdx
-                                    jsr  c64flt.copy_float
+                                    jsr  floats.copy_float
                                 """)
                             }
                             else -> throw AssemblyError("weird target variable type ${target.datatype}")
@@ -401,7 +439,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                                     adc  #<${target.asmVarname}
                                     bcc  +
                                     iny
-+                                   jsr  c64flt.copy_float""")
++                                   jsr  floats.copy_float""")
                             }
                             else -> throw AssemblyError("weird dt")
                         }
@@ -458,13 +496,13 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
 //                    TODO("array[var] ${target.constArrayIndexValue}")
 //                }
                 val index = target.array!!.arrayspec.index
-                asmgen.out("  lda  #<$sourceName |  ldy  #>$sourceName |  jsr  c64flt.push_float")
+                asmgen.out("  lda  #<$sourceName |  ldy  #>$sourceName |  jsr  floats.push_float")
                 asmgen.translateExpression(index)
-                asmgen.out("  lda  #<${target.asmVarname} |  ldy  #>${target.asmVarname} |  jsr  c64flt.pop_float_to_indexed_var")
+                asmgen.out("  lda  #<${target.asmVarname} |  ldy  #>${target.asmVarname} |  jsr  floats.pop_float_to_indexed_var")
             }
             TargetStorageKind.MEMORY -> throw AssemblyError("can't assign float to mem byte")
             TargetStorageKind.REGISTER -> throw AssemblyError("can't assign float to register")
-            TargetStorageKind.STACK -> asmgen.out("  lda  #<$sourceName |  ldy  #>$sourceName |  jsr  c64flt.push_float")
+            TargetStorageKind.STACK -> asmgen.out("  lda  #<$sourceName |  ldy  #>$sourceName |  jsr  floats.push_float")
         }
     }
 
@@ -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  ${target.asmVarname}+1
@@ -751,7 +850,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                             sta  P8ZP_SCRATCH_W1
                             lda  #>${target.asmVarname}
                             sta  P8ZP_SCRATCH_W1+1
-                            jsr  c64flt.set_0_array_float
+                            jsr  floats.set_0_array_float
                         """)
                     }
                 }
@@ -759,7 +858,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                 TargetStorageKind.REGISTER -> throw AssemblyError("can't assign float to register")
                 TargetStorageKind.STACK -> {
                     val floatConst = asmgen.getFloatAsmConst(float)
-                    asmgen.out(" lda  #<$floatConst |  ldy  #>$floatConst |  jsr  c64flt.push_float")
+                    asmgen.out(" lda  #<$floatConst |  ldy  #>$floatConst |  jsr  floats.push_float")
                 }
             }
         } else {
@@ -814,7 +913,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                             sta  P8ZP_SCRATCH_W2
                             lda  #>${arrayVarName}
                             sta  P8ZP_SCRATCH_W2+1
-                            jsr  c64flt.set_array_float
+                            jsr  floats.set_array_float
                         """)
                     }
                 }
@@ -822,7 +921,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                 TargetStorageKind.REGISTER -> throw AssemblyError("can't assign float to register")
                 TargetStorageKind.STACK -> {
                     val floatConst = asmgen.getFloatAsmConst(float)
-                    asmgen.out(" lda  #<$floatConst |  ldy  #>$floatConst |  jsr  c64flt.push_float")
+                    asmgen.out(" lda  #<$floatConst |  ldy  #>$floatConst |  jsr  floats.push_float")
                 }
             }
         }
@@ -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 ${wordtarget.asmVarname} ")
+                }
+                TargetStorageKind.REGISTER -> {
+                    asmgen.loadByteFromPointerIntoA(identifier)
+                    when(wordtarget.register!!) {
+                        RegisterOrPair.AX -> asmgen.out("  ldx  #0")
+                        RegisterOrPair.AY -> asmgen.out("  ldy  #0")
+                        RegisterOrPair.XY -> asmgen.out("  tax |  ldy  #0")
+                        else -> throw AssemblyError("word regs can only be pair")
+                    }
+                }
+                TargetStorageKind.STACK -> {
+                    asmgen.loadByteFromPointerIntoA(identifier)
+                    asmgen.out(" sta  P8ESTACK_LO,x |  lda  #0 |  sta  P8ESTACK_HI,x |  dex")
+                }
+                else -> throw AssemblyError("other types aren't word")
+            }
+        }
+    }
+
     private fun storeByteViaRegisterAInMemoryAddress(ldaInstructionArg: String, memoryAddress: DirectMemoryWrite) {
         val addressExpr = memoryAddress.addressExpression
         val addressLv = addressExpr as? NumericLiteralValue
@@ -957,7 +1117,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                     dex
                     lda  #<$asmArrayvarname
                     ldy  #>$asmArrayvarname
-                    jsr  c64flt.pop_float_to_indexed_var
+                    jsr  floats.pop_float_to_indexed_var
                 """)
             else ->
                 throw AssemblyError("weird array type")

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)
@@ -191,9 +191,8 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                     else -> {
                         println("warning: slow stack evaluation used (1): ${memory.addressExpression::class.simpleName} at ${memory.addressExpression.position}") // TODO optimize...
                         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 +249,9 @@ 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 +310,9 @@ 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 +370,12 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                     asmgen.out("  sta  (P8ZP_SCRATCH_W1),y")
             }
             "*" -> {
-                TODO("mem mul byte litval")
+                if(value in asmgen.optimizedByteMultiplications) {
-                // asmgen.out("  jsr  prog8_lib.mul_byte")  //  the optimized routines should have been checked earlier
+                    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 +450,8 @@ 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
+                // check optimizedByteMultiplications
+                // 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?
+                if(dt == DataType.UBYTE) {
-                TODO("var byte mul litval")
+                    if(value in asmgen.optimizedByteMultiplications) {
-                // asmgen.out("  jsr  prog8_lib.mul_byte")  //  the optimized routines should have been checked earlier
+                        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) {
@@ -669,7 +689,10 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 }
             }
             "*" -> {
-                // TODO what about the optimized mul_5 etc routines?
+                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 {
                         asmgen.out("""
                             lda  $name
                             sta  P8ZP_SCRATCH_W1
@@ -683,6 +706,26 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                             lda  math.multiply_words.result+1
                             sta  $name+1""")
                     }
+                } 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 does this work for signed words? if so the uword/word distinction can be removed altogether
+                        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)
                     throw AssemblyError("division by zero")
@@ -747,13 +790,13 @@ 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 +804,7 @@ 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")
@@ -794,7 +837,7 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
         val valueDt = ident.targetVarDecl(program.namespace)!!.datatype
         when (valueDt) {
             in ByteDatatypes -> {
-                // the other variable is a BYTE type so optimize for that    TODO does this even occur?
+                // the other variable is a BYTE type so optimize for that
                 when (operator) {
                     // note: ** (power) operator requires floats.
                     "+" -> asmgen.out("""
@@ -813,9 +856,22 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                         bcs  +
                         dec  $name+1
 +                           """)
-                    "*" -> TODO("mul word*byte")
+                    "*" -> {
-                    "/" -> TODO("div word/byte")
+                        asmgen.out("""
-                    "%" -> TODO("word remainder byte")
+                            lda  $otherName
+                            sta  P8ZP_SCRATCH_W1
+                            lda  #0
+                            sta  P8ZP_SCRATCH_W1+1
+                            lda  $name
+                            ldy  $name+1
+                            jsr  math.multiply_words
+                            lda  math.multiply_words.result
+                            sta  $name
+                            lda  math.multiply_words.result+1
+                            sta  $name+1""")
+                    }
+                    "/" -> TODO("div wordvar/bytevar")
+                    "%" -> TODO("word remainder bytevar")
                     "<<" -> {
                         asmgen.out("""
                         ldy  $otherName
@@ -843,9 +899,9 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                             bne  -""")
                         }
                     }
-                    "&" -> TODO("bitand word byte")
+                    "&" -> TODO("bitand wordvar bytevar")
-                    "^" -> TODO("bitxor word byte")
+                    "^" -> TODO("bitxor wordvar bytevar")
-                    "|" -> TODO("bitor word byte")
+                    "|" -> TODO("bitor wordvar bytevar")
                     else -> throw AssemblyError("invalid operator for in-place modification $operator")
                 }
             }
@@ -1088,56 +1144,51 @@ 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 ${value.position}") // TODO
         asmgen.translateExpression(value)
+        asmgen.out("  jsr  floats.pop_float_fac1")
         asmgen.saveRegister(CpuRegister.X)
         when (operator) {
             "**" -> {
                 asmgen.out("""
-                    jsr  c64flt.pop_float_fac1
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.CONUPK
+                    jsr  floats.CONUPK
-                    jsr  c64flt.FPWRT
+                    jsr  floats.FPWRT
                 """)
             }
             "+" -> {
                 asmgen.out("""
-                    jsr  c64flt.pop_float_fac1
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.FADD
+                    jsr  floats.FADD
                 """)
             }
             "-" -> {
                 asmgen.out("""
-                    jsr  c64flt.pop_float_fac1
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.FSUB
+                    jsr  floats.FSUB
                 """)
             }
             "*" -> {
                 asmgen.out("""
-                    jsr  c64flt.pop_float_fac1
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.FMULT
+                    jsr  floats.FMULT
                 """)
             }
             "/" -> {
                 asmgen.out("""
-                    jsr  c64flt.pop_float_fac1
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.FDIV
+                    jsr  floats.FDIV
                 """)
             }
             else -> throw AssemblyError("invalid operator for in-place float modification $operator")
         }
-        // store Fac1 back into memory
         asmgen.out("""
             ldx  #<$name
             ldy  #>$name
-            jsr  c64flt.MOVMF
+            jsr  floats.MOVMF
         """)
         asmgen.restoreRegister(CpuRegister.X)
     }
@@ -1154,50 +1205,50 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 asmgen.out("""
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.CONUPK
+                    jsr  floats.CONUPK
                     lda  #<$otherName
                     ldy  #>$otherName
-                    jsr  c64flt.FPWR
+                    jsr  floats.FPWR
                 """)
             }
             "+" -> {
                 asmgen.out("""
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.MOVFM
+                    jsr  floats.MOVFM
                     lda  #<$otherName
                     ldy  #>$otherName
-                    jsr  c64flt.FADD
+                    jsr  floats.FADD
                 """)
             }
             "-" -> {
                 asmgen.out("""
                     lda  #<$otherName
                     ldy  #>$otherName
-                    jsr  c64flt.MOVFM
+                    jsr  floats.MOVFM
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.FSUB
+                    jsr  floats.FSUB
                 """)
             }
             "*" -> {
                 asmgen.out("""
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.MOVFM
+                    jsr  floats.MOVFM
                     lda  #<$otherName
                     ldy  #>$otherName
-                    jsr  c64flt.FMULT
+                    jsr  floats.FMULT
                 """)
             }
             "/" -> {
                 asmgen.out("""
                     lda  #<$otherName
                     ldy  #>$otherName
-                    jsr  c64flt.MOVFM
+                    jsr  floats.MOVFM
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.FDIV
+                    jsr  floats.FDIV
                 """)
             }
             else -> throw AssemblyError("invalid operator for in-place float modification $operator")
@@ -1206,7 +1257,7 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
         asmgen.out("""
             ldx  #<$name
             ldy  #>$name
-            jsr  c64flt.MOVMF
+            jsr  floats.MOVMF
         """)
         asmgen.restoreRegister(CpuRegister.X)
     }
@@ -1219,10 +1270,10 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 asmgen.out("""
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.CONUPK
+                    jsr  floats.CONUPK
                     lda  #<$constValueName
                     ldy  #>$constValueName
-                    jsr  c64flt.FPWR
+                    jsr  floats.FPWR
                 """)
             }
             "+" -> {
@@ -1231,10 +1282,10 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 asmgen.out("""
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.MOVFM
+                    jsr  floats.MOVFM
                     lda  #<$constValueName
                     ldy  #>$constValueName
-                    jsr  c64flt.FADD
+                    jsr  floats.FADD
                 """)
             }
             "-" -> {
@@ -1243,20 +1294,21 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 asmgen.out("""
                     lda  #<$constValueName
                     ldy  #>$constValueName
-                    jsr  c64flt.MOVFM
+                    jsr  floats.MOVFM
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.FSUB
+                    jsr  floats.FSUB
                 """)
             }
             "*" -> {
+                // assume that code optimization is already done on the AST level for special cases such as 0, 1, 2...
                 asmgen.out("""
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.MOVFM
+                    jsr  floats.MOVFM
                     lda  #<$constValueName
                     ldy  #>$constValueName
-                    jsr  c64flt.FMULT
+                    jsr  floats.FMULT
                 """)
             }
             "/" -> {
@@ -1265,10 +1317,10 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 asmgen.out("""
                     lda  #<$constValueName
                     ldy  #>$constValueName
-                    jsr  c64flt.MOVFM
+                    jsr  floats.MOVFM
                     lda  #<$name
                     ldy  #>$name
-                    jsr  c64flt.FDIV
+                    jsr  floats.FDIV
                 """)
             }
             else -> throw AssemblyError("invalid operator for in-place float modification $operator")
@@ -1277,7 +1329,7 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
         asmgen.out("""
             ldx  #<$name
             ldy  #>$name
-            jsr  c64flt.MOVMF
+            jsr  floats.MOVMF
         """)
         asmgen.restoreRegister(CpuRegister.X)
     }
@@ -1285,7 +1337,6 @@ 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 +1346,20 @@ 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  ${target.asmVarname},y")
                                 }
                                 TargetStorageKind.STACK -> {
-                                    if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
+                                    if(CompilationTarget.instance.machine.cpu == CpuType.CPU65c02)
                                         asmgen.out(" stz  P8ESTACK_HI+1,x")
                                     else
                                         asmgen.out(" lda  #0 |  sta  P8ESTACK_HI+1,x")
@@ -1520,11 +1571,11 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                         asmgen.out("""
                             lda  #<${target.asmVarname}
                             ldy  #>${target.asmVarname}
-                            jsr  c64flt.MOVFM
+                            jsr  floats.MOVFM
-                            jsr  c64flt.NEGOP
+                            jsr  floats.NEGOP
                             ldx  #<${target.asmVarname}
                             ldy  #>${target.asmVarname}
-                            jsr  c64flt.MOVMF
+                            jsr  floats.MOVMF
                         """)
                         asmgen.restoreRegister(CpuRegister.X)
                     }

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

@@ -26,14 +26,13 @@ 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)
 
     override fun getFloatRomConst(number: Double): String? = null       // Cx16 has no pulblic ROM float locations
     override fun importLibs(compilerOptions: CompilationOptions, importer: ModuleImporter, program: Program) {
         if (compilerOptions.launcher == LauncherType.BASIC || compilerOptions.output == OutputType.PRG)
-            importer.importLibraryModule(program, "cx16lib")
+            importer.importLibraryModule(program, "syslib")
     }
 
     override fun launchEmulator(programName: String) {
@@ -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

@@ -7,178 +7,6 @@ import prog8.ast.expressions.*
 import prog8.ast.processing.AstWalker
 import prog8.ast.processing.IAstModification
 import prog8.ast.statements.*
-import prog8.compiler.target.CompilationTarget
-
-
-// First thing to do is replace all constant identifiers with their actual value,
-// and the array var initializer values and sizes.
-// This is needed because further constant optimizations depend on those.
-internal class ConstantIdentifierReplacer(private val program: Program, private val errors: ErrorReporter) : AstWalker() {
-    private val noModifications = emptyList<IAstModification>()
-
-    override fun after(identifier: IdentifierReference, parent: Node): Iterable<IAstModification> {
-        // replace identifiers that refer to const value, with the value itself
-        // if it's a simple type and if it's not a left hand side variable
-        if(identifier.parent is AssignTarget)
-            return noModifications
-        var forloop = identifier.parent as? ForLoop
-        if(forloop==null)
-            forloop = identifier.parent.parent as? ForLoop
-        if(forloop!=null && identifier===forloop.loopVar)
-            return noModifications
-
-        val cval = identifier.constValue(program) ?: return noModifications
-        return when (cval.type) {
-            in NumericDatatypes -> listOf(IAstModification.ReplaceNode(identifier, NumericLiteralValue(cval.type, cval.number, identifier.position), identifier.parent))
-            in PassByReferenceDatatypes -> throw FatalAstException("pass-by-reference type should not be considered a constant")
-            else -> noModifications
-        }
-    }
-
-    override fun before(decl: VarDecl, parent: Node): Iterable<IAstModification> {
-        // the initializer value can't refer to the variable itself (recursive definition)
-        // TODO: use call graph for this?
-        if(decl.value?.referencesIdentifiers(decl.name) == true || decl.arraysize?.index?.referencesIdentifiers(decl.name) == true) {
-            errors.err("recursive var declaration", decl.position)
-            return noModifications
-        }
-
-        if(decl.type==VarDeclType.CONST || decl.type==VarDeclType.VAR) {
-            if(decl.isArray){
-                if(decl.arraysize==null) {
-                    // for arrays that have no size specifier (or a non-constant one) attempt to deduce the size
-                    val arrayval = decl.value as? ArrayLiteralValue
-                    if(arrayval!=null) {
-                        return listOf(IAstModification.SetExpression(
-                                { decl.arraysize = ArrayIndex(it, decl.position) },
-                                NumericLiteralValue.optimalInteger(arrayval.value.size, decl.position),
-                                decl
-                        ))
-                    }
-                }
-                else if(decl.arraysize?.constIndex()==null) {
-                    val size = decl.arraysize!!.index.constValue(program)
-                    if(size!=null) {
-                        return listOf(IAstModification.SetExpression(
-                                { decl.arraysize = ArrayIndex(it, decl.position) },
-                                size, decl
-                        ))
-                    }
-                }
-            }
-
-            when(decl.datatype) {
-                DataType.FLOAT -> {
-                    // vardecl: for scalar float vars, promote constant integer initialization values to floats
-                    val litval = decl.value as? NumericLiteralValue
-                    if (litval!=null && litval.type in IntegerDatatypes) {
-                        val newValue = NumericLiteralValue(DataType.FLOAT, litval.number.toDouble(), litval.position)
-                        return listOf(IAstModification.ReplaceNode(decl.value!!, newValue, decl))
-                    }
-                }
-                DataType.ARRAY_UB, DataType.ARRAY_B, DataType.ARRAY_UW, DataType.ARRAY_W -> {
-                    val numericLv = decl.value as? NumericLiteralValue
-                    val rangeExpr = decl.value as? RangeExpr
-                    if(rangeExpr!=null) {
-                        // convert the initializer range expression to an actual array
-                        val declArraySize = decl.arraysize?.constIndex()
-                        if(declArraySize!=null && declArraySize!=rangeExpr.size())
-                            errors.err("range expression size doesn't match declared array size", decl.value?.position!!)
-                        val constRange = rangeExpr.toConstantIntegerRange()
-                        if(constRange!=null) {
-                            val eltType = rangeExpr.inferType(program).typeOrElse(DataType.UBYTE)
-                            val newValue = if(eltType in ByteDatatypes) {
-                                ArrayLiteralValue(InferredTypes.InferredType.known(decl.datatype),
-                                        constRange.map { NumericLiteralValue(eltType, it.toShort(), decl.value!!.position) }.toTypedArray(),
-                                        position = decl.value!!.position)
-                            } else {
-                                ArrayLiteralValue(InferredTypes.InferredType.known(decl.datatype),
-                                        constRange.map { NumericLiteralValue(eltType, it, decl.value!!.position) }.toTypedArray(),
-                                        position = decl.value!!.position)
-                            }
-                            return listOf(IAstModification.ReplaceNode(decl.value!!, newValue, decl))
-                        }
-                    }
-                    if(numericLv!=null && numericLv.type==DataType.FLOAT)
-                        errors.err("arraysize requires only integers here", numericLv.position)
-                    val size = decl.arraysize?.constIndex() ?: return noModifications
-                    if (rangeExpr==null && numericLv!=null) {
-                        // arraysize initializer is empty or a single int, and we know the size; create the arraysize.
-                        val fillvalue = numericLv.number.toInt()
-                        when(decl.datatype){
-                            DataType.ARRAY_UB -> {
-                                if(fillvalue !in 0..255)
-                                    errors.err("ubyte value overflow", numericLv.position)
-                            }
-                            DataType.ARRAY_B -> {
-                                if(fillvalue !in -128..127)
-                                    errors.err("byte value overflow", numericLv.position)
-                            }
-                            DataType.ARRAY_UW -> {
-                                if(fillvalue !in 0..65535)
-                                    errors.err("uword value overflow", numericLv.position)
-                            }
-                            DataType.ARRAY_W -> {
-                                if(fillvalue !in -32768..32767)
-                                    errors.err("word value overflow", numericLv.position)
-                            }
-                            else -> {}
-                        }
-                        // create the array itself, filled with the fillvalue.
-                        val array = Array(size) {fillvalue}.map { NumericLiteralValue(ArrayElementTypes.getValue(decl.datatype), it, numericLv.position) }.toTypedArray<Expression>()
-                        val refValue = ArrayLiteralValue(InferredTypes.InferredType.known(decl.datatype), array, position = numericLv.position)
-                        return listOf(IAstModification.ReplaceNode(decl.value!!, refValue, decl))
-                    }
-                }
-                DataType.ARRAY_F  -> {
-                    val size = decl.arraysize?.constIndex() ?: return noModifications
-                    val litval = decl.value as? NumericLiteralValue
-                    val rangeExpr = decl.value as? RangeExpr
-                    if(rangeExpr!=null) {
-                        // convert the initializer range expression to an actual array of floats
-                        val declArraySize = decl.arraysize?.constIndex()
-                        if(declArraySize!=null && declArraySize!=rangeExpr.size())
-                            errors.err("range expression size doesn't match declared array size", decl.value?.position!!)
-                        val constRange = rangeExpr.toConstantIntegerRange()
-                        if(constRange!=null) {
-                            val newValue = ArrayLiteralValue(InferredTypes.InferredType.known(DataType.ARRAY_F),
-                                        constRange.map { NumericLiteralValue(DataType.FLOAT, it.toDouble(), decl.value!!.position) }.toTypedArray(),
-                                        position = decl.value!!.position)
-                            return listOf(IAstModification.ReplaceNode(decl.value!!, newValue, decl))
-                        }
-                    }
-                    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)
-                            errors.err("float value overflow", litval.position)
-                        else {
-                            // create the array itself, filled with the fillvalue.
-                            val array = Array(size) {fillvalue}.map { NumericLiteralValue(DataType.FLOAT, it, litval.position) }.toTypedArray<Expression>()
-                            val refValue = ArrayLiteralValue(InferredTypes.InferredType.known(DataType.ARRAY_F), array, position = litval.position)
-                            return listOf(IAstModification.ReplaceNode(decl.value!!, refValue, decl))
-                        }
-                    }
-                }
-                else -> {
-                    // nothing to do for this type
-                    // this includes strings and structs
-                }
-            }
-        }
-
-        val declValue = decl.value
-        if(declValue!=null && decl.type==VarDeclType.VAR
-                && declValue is NumericLiteralValue && !declValue.inferType(program).istype(decl.datatype)) {
-            // cast the numeric literal to the appropriate datatype of the variable
-            val cast = declValue.cast(decl.datatype)
-            if(cast.isValid)
-                return listOf(IAstModification.ReplaceNode(decl.value!!, cast.valueOrZero(), decl))
-        }
-
-        return noModifications
-    }
-}
 
 
 internal class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
@@ -216,9 +44,24 @@ internal class ConstantFoldingOptimizer(private val program: Program) : AstWalke
                     else -> throw ExpressionError("can only take negative of int or float", subexpr.position)
                 }
                 "~" -> when (subexpr.type) {
-                    in IntegerDatatypes -> {
+                    DataType.BYTE -> {
                         listOf(IAstModification.ReplaceNode(expr,
-                                NumericLiteralValue.optimalInteger(subexpr.number.toInt().inv(), subexpr.position),
+                                NumericLiteralValue(DataType.BYTE, subexpr.number.toInt().inv(), subexpr.position),
+                                parent))
+                    }
+                    DataType.UBYTE -> {
+                        listOf(IAstModification.ReplaceNode(expr,
+                                NumericLiteralValue(DataType.UBYTE, subexpr.number.toInt().inv() and 255, subexpr.position),
+                                parent))
+                    }
+                    DataType.WORD -> {
+                        listOf(IAstModification.ReplaceNode(expr,
+                                NumericLiteralValue(DataType.WORD, subexpr.number.toInt().inv(), subexpr.position),
+                                parent))
+                    }
+                    DataType.UWORD -> {
+                        listOf(IAstModification.ReplaceNode(expr,
+                                NumericLiteralValue(DataType.UWORD, subexpr.number.toInt().inv() and 65535, subexpr.position),
                                 parent))
                     }
                     else -> throw ExpressionError("can only take bitwise inversion of int", subexpr.position)

compiler/src/prog8/optimizer/ConstantIdentifierReplacer.kt

@@ -0,0 +1,192 @@
+package prog8.optimizer
+
+import prog8.ast.Node
+import prog8.ast.Program
+import prog8.ast.base.*
+import prog8.ast.expressions.*
+import prog8.ast.processing.AstWalker
+import prog8.ast.processing.IAstModification
+import prog8.ast.statements.ArrayIndex
+import prog8.ast.statements.AssignTarget
+import prog8.ast.statements.ForLoop
+import prog8.ast.statements.VarDecl
+import prog8.compiler.target.CompilationTarget
+
+// Fix up the literal value's type to match that of the vardecl
+internal class VarConstantValueTypeAdjuster(private val program: Program, private val errors: ErrorReporter) : AstWalker() {
+    private val noModifications = emptyList<IAstModification>()
+
+    override fun after(decl: VarDecl, parent: Node): Iterable<IAstModification> {
+        val declConstValue = decl.value?.constValue(program)
+        if(declConstValue!=null && (decl.type==VarDeclType.VAR || decl.type==VarDeclType.CONST)
+                && !declConstValue.inferType(program).istype(decl.datatype)) {
+            // cast the numeric literal to the appropriate datatype of the variable
+            val cast = declConstValue.cast(decl.datatype)
+            if(cast.isValid)
+                return listOf(IAstModification.ReplaceNode(decl.value!!, cast.valueOrZero(), decl))
+        }
+        return noModifications
+    }
+}
+
+
+// Replace all constant identifiers with their actual value,
+// and the array var initializer values and sizes.
+// This is needed because further constant optimizations depend on those.
+internal class ConstantIdentifierReplacer(private val program: Program, private val errors: ErrorReporter) : AstWalker() {
+    private val noModifications = emptyList<IAstModification>()
+
+    override fun after(identifier: IdentifierReference, parent: Node): Iterable<IAstModification> {
+        // replace identifiers that refer to const value, with the value itself
+        // if it's a simple type and if it's not a left hand side variable
+        if(identifier.parent is AssignTarget)
+            return noModifications
+        var forloop = identifier.parent as? ForLoop
+        if(forloop==null)
+            forloop = identifier.parent.parent as? ForLoop
+        if(forloop!=null && identifier===forloop.loopVar)
+            return noModifications
+
+        val cval = identifier.constValue(program) ?: return noModifications
+        return when (cval.type) {
+            in NumericDatatypes -> listOf(IAstModification.ReplaceNode(identifier, NumericLiteralValue(cval.type, cval.number, identifier.position), identifier.parent))
+            in PassByReferenceDatatypes -> throw FatalAstException("pass-by-reference type should not be considered a constant")
+            else -> noModifications
+        }
+    }
+
+    override fun before(decl: VarDecl, parent: Node): Iterable<IAstModification> {
+        // the initializer value can't refer to the variable itself (recursive definition)
+        // TODO: use call graph for this?
+        if(decl.value?.referencesIdentifiers(decl.name) == true || decl.arraysize?.index?.referencesIdentifiers(decl.name) == true) {
+            errors.err("recursive var declaration", decl.position)
+            return noModifications
+        }
+
+        if(decl.type== VarDeclType.CONST || decl.type== VarDeclType.VAR) {
+            if(decl.isArray){
+                if(decl.arraysize==null) {
+                    // for arrays that have no size specifier (or a non-constant one) attempt to deduce the size
+                    val arrayval = decl.value as? ArrayLiteralValue
+                    if(arrayval!=null) {
+                        return listOf(IAstModification.SetExpression(
+                                { decl.arraysize = ArrayIndex(it, decl.position) },
+                                NumericLiteralValue.optimalInteger(arrayval.value.size, decl.position),
+                                decl
+                        ))
+                    }
+                }
+                else if(decl.arraysize?.constIndex()==null) {
+                    val size = decl.arraysize!!.index.constValue(program)
+                    if(size!=null) {
+                        return listOf(IAstModification.SetExpression(
+                                { decl.arraysize = ArrayIndex(it, decl.position) },
+                                size, decl
+                        ))
+                    }
+                }
+            }
+
+            when(decl.datatype) {
+                DataType.FLOAT -> {
+                    // vardecl: for scalar float vars, promote constant integer initialization values to floats
+                    val litval = decl.value as? NumericLiteralValue
+                    if (litval!=null && litval.type in IntegerDatatypes) {
+                        val newValue = NumericLiteralValue(DataType.FLOAT, litval.number.toDouble(), litval.position)
+                        return listOf(IAstModification.ReplaceNode(decl.value!!, newValue, decl))
+                    }
+                }
+                DataType.ARRAY_UB, DataType.ARRAY_B, DataType.ARRAY_UW, DataType.ARRAY_W -> {
+                    val numericLv = decl.value as? NumericLiteralValue
+                    val rangeExpr = decl.value as? RangeExpr
+                    if(rangeExpr!=null) {
+                        // convert the initializer range expression to an actual array
+                        val declArraySize = decl.arraysize?.constIndex()
+                        if(declArraySize!=null && declArraySize!=rangeExpr.size())
+                            errors.err("range expression size doesn't match declared array size", decl.value?.position!!)
+                        val constRange = rangeExpr.toConstantIntegerRange()
+                        if(constRange!=null) {
+                            val eltType = rangeExpr.inferType(program).typeOrElse(DataType.UBYTE)
+                            val newValue = if(eltType in ByteDatatypes) {
+                                ArrayLiteralValue(InferredTypes.InferredType.known(decl.datatype),
+                                        constRange.map { NumericLiteralValue(eltType, it.toShort(), decl.value!!.position) }.toTypedArray(),
+                                        position = decl.value!!.position)
+                            } else {
+                                ArrayLiteralValue(InferredTypes.InferredType.known(decl.datatype),
+                                        constRange.map { NumericLiteralValue(eltType, it, decl.value!!.position) }.toTypedArray(),
+                                        position = decl.value!!.position)
+                            }
+                            return listOf(IAstModification.ReplaceNode(decl.value!!, newValue, decl))
+                        }
+                    }
+                    if(numericLv!=null && numericLv.type== DataType.FLOAT)
+                        errors.err("arraysize requires only integers here", numericLv.position)
+                    val size = decl.arraysize?.constIndex() ?: return noModifications
+                    if (rangeExpr==null && numericLv!=null) {
+                        // arraysize initializer is empty or a single int, and we know the size; create the arraysize.
+                        val fillvalue = numericLv.number.toInt()
+                        when(decl.datatype){
+                            DataType.ARRAY_UB -> {
+                                if(fillvalue !in 0..255)
+                                    errors.err("ubyte value overflow", numericLv.position)
+                            }
+                            DataType.ARRAY_B -> {
+                                if(fillvalue !in -128..127)
+                                    errors.err("byte value overflow", numericLv.position)
+                            }
+                            DataType.ARRAY_UW -> {
+                                if(fillvalue !in 0..65535)
+                                    errors.err("uword value overflow", numericLv.position)
+                            }
+                            DataType.ARRAY_W -> {
+                                if(fillvalue !in -32768..32767)
+                                    errors.err("word value overflow", numericLv.position)
+                            }
+                            else -> {}
+                        }
+                        // create the array itself, filled with the fillvalue.
+                        val array = Array(size) {fillvalue}.map { NumericLiteralValue(ArrayElementTypes.getValue(decl.datatype), it, numericLv.position) }.toTypedArray<Expression>()
+                        val refValue = ArrayLiteralValue(InferredTypes.InferredType.known(decl.datatype), array, position = numericLv.position)
+                        return listOf(IAstModification.ReplaceNode(decl.value!!, refValue, decl))
+                    }
+                }
+                DataType.ARRAY_F -> {
+                    val size = decl.arraysize?.constIndex() ?: return noModifications
+                    val litval = decl.value as? NumericLiteralValue
+                    val rangeExpr = decl.value as? RangeExpr
+                    if(rangeExpr!=null) {
+                        // convert the initializer range expression to an actual array of floats
+                        val declArraySize = decl.arraysize?.constIndex()
+                        if(declArraySize!=null && declArraySize!=rangeExpr.size())
+                            errors.err("range expression size doesn't match declared array size", decl.value?.position!!)
+                        val constRange = rangeExpr.toConstantIntegerRange()
+                        if(constRange!=null) {
+                            val newValue = ArrayLiteralValue(InferredTypes.InferredType.known(DataType.ARRAY_F),
+                                    constRange.map { NumericLiteralValue(DataType.FLOAT, it.toDouble(), decl.value!!.position) }.toTypedArray(),
+                                    position = decl.value!!.position)
+                            return listOf(IAstModification.ReplaceNode(decl.value!!, newValue, decl))
+                        }
+                    }
+                    if(rangeExpr==null && litval!=null) {
+                        // arraysize initializer is a single int, and we know the size.
+                        val fillvalue = litval.number.toDouble()
+                        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.
+                            val array = Array(size) {fillvalue}.map { NumericLiteralValue(DataType.FLOAT, it, litval.position) }.toTypedArray<Expression>()
+                            val refValue = ArrayLiteralValue(InferredTypes.InferredType.known(DataType.ARRAY_F), array, position = litval.position)
+                            return listOf(IAstModification.ReplaceNode(decl.value!!, refValue, decl))
+                        }
+                    }
+                }
+                else -> {
+                    // nothing to do for this type
+                    // this includes strings and structs
+                }
+            }
+        }
+
+        return noModifications
+    }
+}

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/Extensions.kt

@@ -5,11 +5,20 @@ import prog8.ast.base.ErrorReporter
 
 
 internal fun Program.constantFold(errors: ErrorReporter) {
+    val valuetypefixer = VarConstantValueTypeAdjuster(this, errors)
+    valuetypefixer.visit(this)
+    if(errors.isEmpty()) {
+        valuetypefixer.applyModifications()
+
         val replacer = ConstantIdentifierReplacer(this, errors)
         replacer.visit(this)
         if (errors.isEmpty()) {
             replacer.applyModifications()
 
+            valuetypefixer.visit(this)
+            if(errors.isEmpty()) {
+                valuetypefixer.applyModifications()
+
                 val optimizer = ConstantFoldingOptimizer(this)
                 optimizer.visit(this)
                 while (errors.isEmpty() && optimizer.applyModifications() > 0) {
@@ -21,6 +30,8 @@ internal fun Program.constantFold(errors: ErrorReporter) {
                     replacer.applyModifications()
                 }
             }
+        }
+    }
 
     if(errors.isEmpty())
         modules.forEach { it.linkParents(namespace) }   // re-link in final configuration

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/src/prog8/parser/ModuleParsing.kt

@@ -9,6 +9,7 @@ import prog8.ast.base.SyntaxError
 import prog8.ast.base.checkImportedValid
 import prog8.ast.statements.Directive
 import prog8.ast.statements.DirectiveArg
+import prog8.compiler.target.CompilationTarget
 import prog8.pathFrom
 import java.io.InputStream
 import java.nio.file.Files
@@ -19,14 +20,6 @@ import java.nio.file.Paths
 internal class ParsingFailedError(override var message: String) : Exception(message)
 
 
-private class LexerErrorListener: BaseErrorListener() {
-    var  numberOfErrors: Int = 0
-    override fun syntaxError(p0: Recognizer<*, *>?, p1: Any?, p2: Int, p3: Int, p4: String?, p5: RecognitionException?) {
-        numberOfErrors++
-    }
-}
-
-
 internal class CustomLexer(val modulePath: Path, input: CharStream?) : prog8Lexer(input)
 
 
@@ -60,13 +53,28 @@ internal class ModuleImporter {
         return executeImportDirective(program, import, Paths.get(""))
     }
 
+    private class MyErrorListener: ConsoleErrorListener() {
+        var  numberOfErrors: Int = 0
+        override fun syntaxError(recognizer: Recognizer<*, *>?, offendingSymbol: Any?, line: Int, charPositionInLine: Int, msg: String, e: RecognitionException?) {
+            numberOfErrors++
+            when (recognizer) {
+                is CustomLexer -> System.err.println("${recognizer.modulePath}:$line:$charPositionInLine: $msg")
+                is prog8Parser -> System.err.println("${recognizer.inputStream.sourceName}:$line:$charPositionInLine: $msg")
+                else -> System.err.println("$line:$charPositionInLine $msg")
+            }
+        }
+    }
+
     private fun importModule(program: Program, stream: CharStream, modulePath: Path, isLibrary: Boolean): Module {
         val moduleName = moduleName(modulePath.fileName)
         val lexer = CustomLexer(modulePath, stream)
-        val lexerErrors = LexerErrorListener()
+        lexer.removeErrorListeners()
+        val lexerErrors = MyErrorListener()
         lexer.addErrorListener(lexerErrors)
         val tokens = CommentHandlingTokenStream(lexer)
         val parser = prog8Parser(tokens)
+        parser.removeErrorListeners()
+        parser.addErrorListener(MyErrorListener())
         val parseTree = parser.module()
         val numberOfErrors = parser.numberOfSyntaxErrors + lexerErrors.numberOfErrors
         if(numberOfErrors > 0)
@@ -141,6 +149,10 @@ internal class ModuleImporter {
     }
 
     private fun tryGetEmbeddedResource(name: String): InputStream? {
+        val target = CompilationTarget.instance.name
+        val targetSpecific = object{}.javaClass.getResourceAsStream("/prog8lib/$target/$name")
+        if(targetSpecific!=null)
+            return targetSpecific
         return object{}.javaClass.getResourceAsStream("/prog8lib/$name")
     }
 }

compiler/test/UnitTests.kt

@@ -185,7 +185,7 @@ class TestC64Zeropage {
         val zp1 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), true))
         assertEquals(16, zp1.available())
         val zp2 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FLOATSAFE, emptyList(), false))
-        assertEquals(91, zp2.available())
+        assertEquals(89, zp2.available())
         val zp3 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.KERNALSAFE, emptyList(), false))
         assertEquals(125, zp3.available())
         val zp4 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false))

docs/source/building.rst

@@ -149,10 +149,10 @@ If your running program hits one of the breakpoints, Vice will halt execution an
 Troubleshooting
 ---------------
 
-Getting an assembler error about undefined symbols such as ``not defined 'c64flt'``?
+Getting an assembler error about undefined symbols such as ``not defined 'floats'``?
-This happens when your program uses floating point values, and you forgot to import ``c64flt`` library.
+This happens when your program uses floating point values, and you forgot to import ``floats`` library.
 If you use floating points, the compiler needs routines from that library.
-Fix it by adding an ``%import c64flt``.
+Fix it by adding an ``%import floats``.
 
 
 Examples

docs/source/index.rst

@@ -38,21 +38,22 @@ This software is licensed under the GNU GPL 3.0, see https://www.gnu.org/license
     :alt: Fully playable tetris clone
 
 
-Code examples
+Code example
--------------
+------------
 
 This code calculates prime numbers using the Sieve of Eratosthenes algorithm::
 
-    %import c64textio
+    %import textio
     %zeropage basicsafe
 
     main {
-
         ubyte[256] sieve
-        ubyte candidate_prime = 2
+        ubyte candidate_prime = 2       ; is increased in the loop
 
         sub start() {
-            memset(sieve, 256, false)   ; clear the sieve
+            ; clear the sieve, to reset starting situation on subsequent runs
+            memset(sieve, 256, false)
+            ; calculate primes
             txt.print("prime numbers up to 255:\n\n")
             ubyte amount=0
             repeat {
@@ -63,18 +64,19 @@ This code calculates prime numbers using the Sieve of Eratosthenes algorithm::
                 txt.print(", ")
                 amount++
             }
-            c64.CHROUT('\n')
+            txt.chrout('\n')
             txt.print("number of primes (expected 54): ")
             txt.print_ub(amount)
-            c64.CHROUT('\n')
+            txt.chrout('\n')
         }
 
         sub find_next_prime() -> ubyte {
             while sieve[candidate_prime] {
                 candidate_prime++
                 if candidate_prime==0
-                    return 0        ; we wrapped; no more primes available
+                    return 0        ; we wrapped; no more primes available in the sieve
             }
+
             ; found next one, mark the multiples and return it.
             sieve[candidate_prime] = true
             uword multiple = candidate_prime
@@ -88,7 +90,6 @@ This code calculates prime numbers using the Sieve of Eratosthenes algorithm::
     }
 
 
-
 when compiled an ran on a C-64 you get this:
 
 .. image:: _static/primes_example.png
@@ -96,43 +97,6 @@ when compiled an ran on a C-64 you get this:
     :alt: result when run on C-64
 
 
-The following programs shows a use of the high level ``struct`` type::
-
-
-    %import c64textio
-    %zeropage basicsafe
-
-    main {
-
-        struct Color {
-            ubyte red
-            ubyte green
-            ubyte blue
-        }
-
-        sub start() {
-
-            Color purple = [255, 0, 255]
-
-            Color other
-
-            other = purple
-            other.red /= 2
-            other.green = 10 + other.green / 2
-            other.blue = 99
-
-            txt.print_ub(other.red)
-            c64.CHROUT(',')
-            txt.print_ub(other.green)
-            c64.CHROUT(',')
-            txt.print_ub(other.blue)
-            c64.CHROUT('\n')
-        }
-    }
-
-
-when compiled and ran, it prints  ``127,10,99`` on the screen.
-
 
 Design principles and features
 ------------------------------
@@ -162,6 +126,7 @@ Design principles and features
   the ability to easily write embedded assembly code directly in the program source code.
 - There are many built-in functions, such as ``sin``, ``cos``, ``rnd``, ``abs``, ``min``, ``max``, ``sqrt``, ``msb``, ``rol``, ``ror``, ``swap``, ``memset``, ``memcopy``, ``substr``, ``sort`` and ``reverse`` (and others)
 - Assembling the generated code into a program wil be done by an external cross-assembler tool.
+- If you only use standard kernel and prog8 library routines, it is possible to compile the *exact same program* for both machines (just change the compiler target flag)!
 
 
 .. _requirements:

docs/source/programming.rst

@@ -226,7 +226,7 @@ This is because routines in the C-64 BASIC and KERNAL ROMs are used for that.
 So floating point operations will only work if the C-64 BASIC ROM (and KERNAL ROM)
 are banked in.
 
-Also your code needs to import the ``c64flt`` library to enable floating point support
+Also your code needs to import the ``floats`` library to enable floating point support
 in the compiler, and to gain access to the floating point routines.
 (this library contains the directive to enable floating points, you don't have
 to worry about this yourself)
@@ -578,25 +578,24 @@ within parentheses will be evaluated first. So ``(4 + 8) * 2`` is 24 and not 20,
 and ``(true or false) and false`` is false instead of true.
 
 .. attention::
-    **calculations keep their datatype:**
+    **calculations keep their datatype even if the target variable is larger:**
     When you do calculations on a BYTE type, the result will remain a BYTE.
     When you do calculations on a WORD type, the result will remain a WORD.
     For instance::
 
         byte b = 44
         word w = b*55   ; the result will be 116! (even though the target variable is a word)
-        w *= 999        ; the result will be -15188  (the multiplication stays within a word)
+        w *= 999        ; the result will be -15188  (the multiplication stays within a word, but overflows)
 
-    The compiler will NOT give a warning about this! It's doing this for
+    *The compiler does NOT warn about this!* It's doing this for
     performance reasons - so you won't get sudden 16 bit (or even float)
     calculations where you needed only simple fast byte arithmetic.
     If you do need the extended resulting value, cast at least one of the
-    operands of an operator to the larger datatype. For example::
+    operands explicitly to the larger datatype. For example::
 
         byte b = 44
-        word w = b*55.w     ; the result will be 2420
+        w = (b as word)*55
-        w = (b as word)*55  ; same result
+        w = b*(55 as word)
-
 
 
 
@@ -803,7 +802,7 @@ memset(address, numbytes, bytevalue)
     Efficiently set a part of memory to the given (u)byte value.
     But the most efficient will always be to write a specialized fill routine in assembly yourself!
     Note that for clearing the character screen, very fast specialized subroutines are
-    available in the ``screen`` block (part of the ``c64textio`` or ``cx16textio`` modules)
+    available in the ``txt`` block (part of the ``textio`` module)
 
 memsetw(address, numwords, wordvalue)
     Efficiently set a part of memory to the given (u)word value.

docs/source/syntaxreference.rst

@@ -33,6 +33,13 @@ This makes it easier to understand and relate the generated code. Examples::
 Directives
 -----------
 
+.. data:: %target <target>
+
+    Level: module.
+    Global setting, specifies that this module can only work for the given compiler target.
+    If compiled with a different target, compilation is aborted with an error message.
+
+
 .. data:: %output <type>
 
 	Level: module.
@@ -289,7 +296,8 @@ of something with an operand starting with 1 or 0, you'll have to add a space in
 - When an integer value ranges from 256..65535 the compiler sees it as a ``uword``.  For -32768..32767 it's a ``word``.
 - When a hex number has 3 or 4 digits, for example ``$0004``, it is seen as a ``word`` otherwise as a ``byte``.
 - When a binary number has 9 to 16 digits, for example ``%1100110011``, it is seen as a ``word`` otherwise as a ``byte``.
-- You can force a byte value into a word value by adding the ``.w`` datatype suffix to the number: ``$2a.w`` is equivalent to ``$002a``.
+- If the number fits in a byte but you really require it as a word value, you'll have to explicitly cast it: ``60 as uword``
+  or you can use the full word hexadecimal notation ``$003c``.
 
 
 Data type conversion
@@ -514,17 +522,19 @@ 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.
+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 like this::
+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  $62
+            ldy  #0
-            eor  #$ff
+    _loop   sta  c64.Screen,y
-            asl  a
+            sta  c64.Screen+$0100,y
-            lda  #0
+            sta  c64.Screen+$0200,y
-            ldx  #$a0
+            sta  c64.Screen+$02e8,y
-            jmp  $bc4f
+            iny
+            bne  _loop
+            rts
             }}
     }
 

docs/source/targetsystem.rst

@@ -9,13 +9,16 @@ Prog8 targets the following hardware:
 - optional use of memory mapped I/O registers
 - optional use of system ROM routines
 
-Currently there are two machines that are supported as compiler target (via the ``-target`` compiler argument):
+Currently there are two machines that are supported as compiler target (selectable via the ``-target`` compiler argument):
 
 - 'c64': the well-known Commodore-64, premium support
 - 'cx16': the `CommanderX16 <https://www.commanderx16.com/>`_ a project from the 8-Bit Guy. Support for this is still experimental.
 
 This chapter explains the relevant system details of these machines.
 
+.. note::
+    If you only use standard kernel and prog8 library routines, it is possible to compile the *exact same program* for both machines (just change the compiler target flag)!
+
 
 Memory Model
 ============

docs/source/todo.rst

@@ -2,14 +2,20 @@
 TODO
 ====
 
-- optimize assignment codegeneration
+- get rid of all other TODO's in the code ;-)
-- get rid of all TODO's ;-)
+- move the 	ldx  #$ff | clc | cld  from the startup logic into the start() function as first instructions
+- add an %option that omits the 'system-init' code at the start. Useful to create separate standalone routines that shouldn't re-init the whole machine every time they're called
+- line-circle-gfx examples are now a few hundred bytes larger than before. Why is that, can it be fixed?
+- until condition should be able to refer to variables defined IN the do-until block itself.
+- add support? example? for processing arguments to a sys call : sys 999, 1, 2, "aaa"
+- make it possible for array literals to not only contain compile time constants
+- further optimize assignment codegeneration
+- 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
-
+- use VIC banking to move up the graphics bitmap memory location. Don't move it under the ROM though as that would require IRQ disabling and memory bank swapping for every bitmap manipulation
+- add some primitives/support/examples for using custom char sets, copying the default charset.
 
 More optimizations
 ^^^^^^^^^^^^^^^^^^
@@ -18,14 +24,13 @@ Add more compiler optimizations to the existing ones.
 
 - more targeted optimizations for assigment asm code, such as the following:
 - subroutine calling convention? like: 1 byte arg -> pass in A, 2 bytes -> pass in A+Y, return value likewise.
-- remove unreachable code after an exit(), return or goto
+- can such parameter passing to subroutines be optimized to avoid copying?
 - add a compiler option to not include variable initialization code (useful if the program is expected to run only once, such as a game)
   the program will then rely solely on the values as they are in memory at the time of program startup.
 - Also some library routines and code patterns could perhaps be optimized further
-- can the parameter passing to subroutines be optimized to avoid copying?
 - more optimizations on the language AST level
 - more optimizations on the final assembly source level
-- note: abandoned subroutine inlining because of problems referencing non-local stuff. Can't move everything around.
+- note: subroutine inlining is abandoned because of problems referencing non-local stuff. Can't move everything around.
 
 
 Eval stack redesign? (lot of work)

examples/arithmetic/aggregates.p8

@@ -1,5 +1,5 @@
-%import c64flt
+%import floats
-%import c64textio
+%import textio
 %zeropage basicsafe
 
 main {

examples/arithmetic/bitshift.p8

@@ -1,4 +1,4 @@
-%import c64textio
+%import textio
 %zeropage basicsafe
 
 

examples/arithmetic/div.p8

@@ -1,5 +1,5 @@
-%import c64flt
+%import floats
-%import c64textio
+%import textio
 %zeropage basicsafe
 
 main {
@@ -93,11 +93,11 @@ main {
             txt.print("err! ")
 
         txt.print("float ")
-        c64flt.print_f(a1)
+        floats.print_f(a1)
         txt.print(" / ")
-        c64flt.print_f(a2)
+        floats.print_f(a2)
         txt.print(" = ")
-        c64flt.print_f(r)
+        floats.print_f(r)
         c64.CHROUT('\n')
     }
 }

examples/arithmetic/minus.p8

@@ -1,5 +1,5 @@
-%import c64flt
+%import floats
-%import c64textio
+%import textio
 %zeropage basicsafe
 
 main {
@@ -101,11 +101,11 @@ main {
             txt.print("err! ")
 
         txt.print("float ")
-        c64flt.print_f(a1)
+        floats.print_f(a1)
         txt.print(" - ")
-        c64flt.print_f(a2)
+        floats.print_f(a2)
         txt.print(" = ")
-        c64flt.print_f(r)
+        floats.print_f(r)
         c64.CHROUT('\n')
     }
 }

examples/arithmetic/mult.p8

@@ -1,5 +1,5 @@
-%import c64flt
+%import floats
-%import c64textio
+%import textio
 %zeropage basicsafe
 
 main {
@@ -95,11 +95,11 @@ main {
             txt.print("err! ")
 
         txt.print("float ")
-        c64flt.print_f(a1)
+        floats.print_f(a1)
         txt.print(" * ")
-        c64flt.print_f(a2)
+        floats.print_f(a2)
         txt.print(" = ")
-        c64flt.print_f(r)
+        floats.print_f(r)
         c64.CHROUT('\n')
     }
 }

examples/arithmetic/plus.p8

@@ -1,5 +1,5 @@
-%import c64flt
+%import floats
-%import c64textio
+%import textio
 %zeropage basicsafe
 
 main {
@@ -99,11 +99,11 @@ main {
             txt.print("err! ")
 
         txt.print("float ")
-        c64flt.print_f(a1)
+        floats.print_f(a1)
         txt.print(" + ")
-        c64flt.print_f(a2)
+        floats.print_f(a2)
         txt.print(" = ")
-        c64flt.print_f(r)
+        floats.print_f(r)
         c64.CHROUT('\n')
     }
 }

examples/arithmetic/postincrdecr.p8

@@ -1,5 +1,5 @@
-%import c64flt
+%import floats
-%import c64textio
+%import textio
 %zeropage basicsafe
 
 main {
@@ -132,9 +132,9 @@ main {
         else
             txt.print("err! ")
         txt.print(" float ")
-        c64flt.print_f(value)
+        floats.print_f(value)
         c64.CHROUT(',')
-        c64flt.print_f(expected)
+        floats.print_f(expected)
         c64.CHROUT('\n')
     }
 }

examples/arithmetic/remainder.p8

@@ -1,4 +1,4 @@
-%import c64textio
+%import textio
 %zeropage basicsafe
 
 main {

examples/arithmetic/sgn.p8

@@ -1,5 +1,5 @@
-%import c64flt
+%import floats
-%import c64textio
+%import textio
 %zeropage basicsafe
 
 main {

examples/balloonflight.p8

@@ -1,5 +1,6 @@
-%import c64lib
+%target c64
-%import c64textio
+%import syslib
+%import textio
 %zeropage basicsafe
 
 main {
@@ -43,7 +44,7 @@ main {
             }
 
             perform_scroll = false
-            txt.scroll_left_full(true)
+            txt.scroll_left(true)
             if c64.RASTER & 1
                 c64.SPXY[1] ++
             else

examples/bdmusic-irq.p8

@@ -1,5 +1,6 @@
-%import c64lib
+%target c64
-%import c64textio
+%import syslib
+%import textio
 %zeropage basicsafe
 
 main {

examples/bdmusic.p8

@@ -1,4 +1,6 @@
-%import c64textio
+%target c64
+%import textio
+%import syslib
 
 main {
 

examples/cmp/comparison_ifs_byte.p8

@@ -1,6 +1,7 @@
-%import c64textio
+%import textio
 %zeropage basicsafe
 
+; Note: this program is compatible with C64 and CX16.
 
 main {
 

examples/cmp/comparison_ifs_float.p8

@@ -1,7 +1,9 @@
-%import c64textio
+%import textio
-%import c64flt
+%import floats
 %zeropage basicsafe
 
+; Note: this program is compatible with C64 and CX16.
+
 main {
 
     sub start()  {

examples/cmp/comparison_ifs_ubyte.p8

@@ -1,7 +1,9 @@
-%import c64textio
+%import textio
 %zeropage basicsafe
 
 
+; Note: this program is compatible with C64 and CX16.
+
 main {
 
     sub start()  {

examples/cmp/comparison_ifs_uword.p8

@@ -1,6 +1,7 @@
-%import c64textio
+%import textio
 %zeropage basicsafe
 
+; Note: this program is compatible with C64 and CX16.
 
 main {
 

examples/cmp/comparison_ifs_word.p8

@@ -1,6 +1,7 @@
-%import c64textio
+%import textio
 %zeropage basicsafe
 
+; Note: this program is compatible with C64 and CX16.
 
 main {
 

examples/cmp/comparisons_byte.p8

@@ -1,6 +1,8 @@
-%import c64textio
+%import textio
 %zeropage basicsafe
 
+; Note: this program is compatible with C64 and CX16.
+
 main {
 
     sub start()  {

examples/cmp/comparisons_float.p8

@@ -1,7 +1,9 @@
-%import c64textio
+%import textio
-%import c64flt
+%import floats
 %zeropage basicsafe
 
+; Note: this program is compatible with C64 and CX16.
+
 main {
 
     sub start()  {

examples/cmp/comparisons_ubyte.p8

@@ -1,6 +1,8 @@
-%import c64textio
+%import textio
 %zeropage basicsafe
 
+; Note: this program is compatible with C64 and CX16.
+
 main {
 
     sub start()  {

examples/cmp/comparisons_uword.p8

@@ -1,6 +1,8 @@
-%import c64textio
+%import textio
 %zeropage basicsafe
 
+; Note: this program is compatible with C64 and CX16.
+
 main {
 
     sub start()  {

examples/cmp/comparisons_word.p8

@@ -1,6 +1,8 @@
-%import c64textio
+%import textio
 %zeropage basicsafe
 
+; Note: this program is compatible with C64 and CX16.
+
 main {
 
     sub start()  {