fix floats.parse_f() to use new kernal routine address for VAL

gives error message if it detects issues f.ex. with new kernal version that moves the routine

compiler/res/prog8lib/cx16/floats.p8

@@ -139,13 +139,27 @@ asmsub  FREADUY (ubyte value @Y) {
 asmsub parse_f(str value @AY) -> float @FAC1 {
     ; -- parse a string value of a number to float in FAC1
     ;    warning: uses an internal BASIC routine that may be rom version dependent
-    ;    ($ddf2 is inside the routine for VAL at $ddef)
+    ;    ($deb6 is inside the routine for VAL at $deb3)  See basic.sym from x16-rom
+    ;    If at any time in the future the official VAL_1() routine from the kernal starts working, we should use that instead!
     %asm {{
         sta  $a9
         sty  $aa
         jsr  prog8_lib.strlen
+        lda  $deb6
+        cmp  #$d0   ; sanity check for kernal routine correct
+        bne  +      ;
         tya
-        jmp  $ddf2
+        jmp  $deb6   ; kernal version dependent :(
++       ; print error message if routine is borked in kernal, and exit program
+        ldy  #0
+-       lda  _msg,y
+        beq  +
+        jsr  cbm.CHROUT
+        iny
+        bne  -
++       jmp  sys.exit
+
+_msg    .text 13,"?val kaputt",13,0
     }}
 }
 

compiler/res/prog8lib/cx16/syslib.p8

@@ -1458,7 +1458,7 @@ save_SCRATCH_ZPWORD2	.word  0
         }}
     }
 
-    inline asmsub exit(ubyte returnvalue @A) {
+    asmsub exit(ubyte returnvalue @A) {
         ; -- immediately exit the program with a return code in the A register
         %asm {{
             jsr  cbm.CLRCHN		; reset i/o channels

docs/source/todo.rst

@@ -2,10 +2,6 @@
 TODO
 ====
 
-- fix floats.parse_f -  did the routine move in kernal rom?
-    VAL at $ddef. Looks like it's now $deb3  (see basic.sym from x16-rom)
-    so assuming its internals hasen't changed you need to change the $ddf2 to $deb6?
-
 - [on branch: shortcircuit] investigate McCarthy evaluation again? this may also reduce code size perhaps for things like if a>4 or a<2 ....
 
 ...

examples/test.p8

@@ -4,39 +4,12 @@
 
 main {
     sub start() {
-        ubyte[100] storage = 0
+        txt.print("enter number: ")
-
+        str buffer = "???????????????????????????"
-        txt.print("sizeof float = ")
+        void txt.input_chars(buffer)
-        txt.print_ub(sizeof(0.0))
+        float value = floats.parse_f(buffer)
-        txt.print("\nsizeof word = ")
-        txt.print_ub(sizeof($0000))
-        txt.print("\nsizeof byte = ")
-        txt.print_ub(sizeof($00))
-        txt.print("\nsizeof bool = ")
-        txt.print_ub(sizeof(true))
         txt.nl()
-
+        floats.print_f(value)
-        poke(&storage+10, 123)
-        pokew(&storage+11, 54321)
-        txt.print_ub(peek(&storage+10))
-        txt.spc()
-        txt.print_uw(peekw(&storage+11))
-        txt.nl()
-
-        pokef(&storage+10, 3.14)
-        pokef($4000, 123.456)
-        floats.print_f(peekf(&storage+10))
-        txt.nl()
-        floats.print_f(peekf($4000))
-        txt.nl()
-        pokef(&storage+10, 3.1415927)
-        floats.print_f(peekf(&storage+10))
-        txt.nl()
-
-        for cx16.r2L in 0 to 20 {
-            txt.print_ubhex(storage[cx16.r2L], false)
-            txt.spc()
-        }
         txt.nl()
     }
 }