Better code check before compile?

[pedja089]

From manual

2.18 Line-Concatenation ( : )
Multiple commands may be written on a single line using colon characters to tell PBP where to insert a "virtual" line break.

[HenrikOlsson]

I agree with peja089, the colon acts as a virtual line break.
If more than a signle statement should execute as a result of the evaluation use IF-THEN-ENDIF.

With that said... In section 5.35 of the manual they actually DO show the following as a valid example:

IF B0 <> 10 THEN B0 = B0 + 1: B1 = B1 - 1

And in fact, when compiling the followong code for a 16F628 both the Test1 and Test2 routines results in the same assembly code

A VAR BYTE
B VAR BYTE
C VAR BYTE
D VAR BYTE

Test1:
IF A = 12 THEN B = 2 : C = 3 :  D = 4

Test2:
IF A = 12 THEN
  B=2
  C=3
  D=4
ENDIF

Here's the resulting assembly (both versions looks identical to me (wasn't really expecting that to be honest):

_Test1
         clrwdt
         movf    _A,  W
         sublw   00Ch
         btfss   STATUS, Z
         goto    L00001
         movlw   low (002h)
         movwf   _B
         movlw   low (003h)
         movwf   _C
         movlw   low (004h)
         movwf   _D
L00001
_Test2
        clrwdt
        movf    _A,  W
        sublw   00Ch
        btfss   STATUS, Z
        goto    L00003
        movlw   low (002h)
        movwf   _B
        movlw   low (003h)
        movwf   _C
        movlw   low (004h)
        movwf   _D
L00003

That got me curious so I compiled both version of the original code

ticker VAR BYTE
ard VAR BYTE
atrig VAR BYTE

Test1:
if ticker>150 and ard<140 and atrig=0 then ticker=1: atrig=1

Test2:
if ticker>150 and ard<140 and atrig=0 then
  ticker=1
  atrig=1
ENDIF

And the resulting assebly (less library code) looks like this:

_Test1
        movf    _ticker, W
        movwf   R0
        movlw   low (096h)
        call    CMPGTB
        movwf   T1
        movf    _ard, W
        movwf   R0
        movlw   low (08Ch)
        call    CMPLTB
        movwf   T2
        movf    T1,  W
        movwf   FSR
        movf    T2,  W
        call    LAND
        movwf   T2
        movwf   T2   + 1
        movf    _atrig, W
        sublw   000h
        btfss   STATUS, Z
        movlw   -1
        xorlw   0ffh
        movwf   T3
        movf    T2,  W
        iorwf   T2  + 1, W
        movwf   FSR
        movf    T3,  W
        call    LAND
        movwf   T3
        movwf   T3   + 1
        clrwdt
        movf    T3,  W
        iorwf   (T3)  + 1, W
        btfsc   STATUS, Z
        goto    L00001
        movlw   low (001h)
        movwf   _ticker
        movlw   low (001h)
        movwf   _atrig
L00001
_Test2
        movf    _ticker, W
        movwf   R0
        movlw   low (096h)
        call    CMPGTB
        movwf   T1
        movf    _ard, W
        movwf   R0
        movlw   low (08Ch)
        call    CMPLTB
        movwf   T2
        movf    T1,  W
        movwf   FSR
        movf    T2,  W
        call    LAND
        movwf   T2
        movwf   T2   + 1
        movf    _atrig, W
        sublw   000h
        btfss   STATUS, Z
        movlw   -1
        xorlw   0ffh
        movwf   T3
        movf    T2,  W
        iorwf   T2  + 1, W
        movwf   FSR
        movf    T3,  W
        call    LAND
        movwf   T3
        movwf   T3   + 1
        clrwdt
        movf    T3,  W
        iorwf   (T3)  + 1, W
        btfsc   STATUS, Z
        goto    L00003
        movlw   low (001h)
        movwf   _ticker
        movlw   low (001h)
        movwf   _atrig
L00003

As far as I can see they are identical, meaning atrig will only get set to 1 if the evaulation is true. So I don't know what might be happening in CuriousOnes case because as far as I can see it IS working as CuriousOne expects - which I did NOT expect.

/Henrik.

[pedja089]

... which I did NOT expect.
/Henrik.

Same here