Logical operators bugged? (*solved*)

[britlion]

I can't see why this doesn't work:

Code:

dim a as uByte

FUNCTION get (capitalise as byte) as uByte
    DIM lastK AS uByte AT 23560: REM LAST_K System VAR
    lastK=0
    DO LOOP until lastK <> 0 : REM Wait for a keypress

    IF capitalise >0 AND lastK >= 97 AND lastK <= 122 THEN:
            LET lastK = lastK - 32
    END IF

    IF capitalise <0 AND lastK >=65 AND lastK <= 90 THEN:
        LET lastK=lastK + 32          
    END IF    
    
    RETURN lastK
END FUNCTION

let a=get(1)

print a,chr$(a)

Without fail this doesn't capitalise the key I pressed. get(-1) should only produce lower case (a-z), get(0) should produce what's pressed (a-Z) and get(1) should produce UPPER case (A-Z). It doesn't seem to be changing it. I've proved the IF works - I think the AND is failing.

I tried

Code:

IF ((capitalise >0 AND lastK >= 97) AND lastK <= 122) THEN:

in order to see if' it's something in the multiple AND options. Any one of the three conditions, without the AND seems to come out true.

Are logical functions misbehaving?

[britlion]

Code:

dim i as uByte

FUNCTION CONTRAST (colour as uByte) as uByte
    IF colour <=3 then return 7: END IF
    return 0
END FUNCTION

FOR i = 0 to 7
PRINT "SENT ";i;" GOT ";CONTRAST(i)
NEXT i

This comes back with:

Code:

Sent 0 Got 0     Should be     Sent 0 got 7
Sent 1 Got 0     Should be     Sent 1 got 7
Sent 2 Got 0     Should be     Sent 2 got 7
Sent 3 Got 7     Should be     Sent 3 got 7
Sent 4 Got 7     Should be     Sent 4 got 0
Sent 5 Got 7     Should be     Sent 5 got 0
Sent 6 Got 7     Should be     Sent 6 got 0
Sent 7 Got 7     Should be     Sent 7 got 0

Okay... with "<=3" at least 3 should be the same as 2...yes? And why do 0,1,2 get a return of 0 when they should get a return of 7?

Things get weirder:

Assuming that it was some weird return issue, I decided to return a variable instead:

Code:

dim i as uByte

FUNCTION CONTRAST (colour as uByte) as uByte
    DIM rt as uByte
    IF colour <=3 then rt=7: END IF
    IF colour >=4 then rt=0: END IF    
    return rt
END FUNCTION

FOR i = 0 to 7
PRINT "SENT ";i;" GOT ";CONTRAST(i)
NEXT i

This gave me:

This comes back with:

Code:

Sent 0 Got 0
Sent 1 Got 0
Sent 2 Got 0
Sent 3 Got 7
Sent 4 Got 0
Sent 5 Got 0
Sent 6 Got 0
Sent 7 Got 0

I mean. SRSLY? My opinion of what that code does is clearly disagreeing with zxb's version..

Perhaps it's because we didn't initialise rt?

Code:

dim i as uByte

FUNCTION CONTRAST (colour as uByte) as uByte
    DIM rt as uByte
    rt=9
    IF colour <=3 then rt=7: END IF
    IF colour >=4 then rt=0: END IF    
    return rt
END FUNCTION

FOR i = 0 to 7
PRINT "SENT ";i;" GOT ";CONTRAST(i)
NEXT i

Gives us a little bit of a clearer picture:

Code:

Sent 0 Got 9
Sent 1 Got 9
Sent 2 Got 9
Sent 3 Got 7
Sent 4 Got 0
Sent 5 Got 0
Sent 6 Got 0
Sent 7 Got 0

So. We seem to have a complete logic failure.

0 <= 3 is FALSE and under certain circumstances 7<=3 is TRUE.

I am so going to bed. I've programmed myself into something psychedelic.

Incidentally:

Code:

FUNCTION CONTRAST (colour as uByte) as uByte
    DIM rt as uByte
    IF colour <4 then LET rt = 7: END IF
    IF colour >3 then let rt = 0: END IF
    return rt
END FUNCTION

Seems to be working.

I suppose just asking for PAPER 9 and INK 9 to work is out of the question?

[boriel]

Code:

dim i as uByte

FUNCTION CONTRAST (colour as uByte) as uByte
    IF colour <=3 then return 7: END IF
    return 0
END FUNCTION

FOR i = 0 to 7
PRINT "SENT ";i;" GOT ";CONTRAST(i)
NEXT i

This comes back with:

Code:

Sent 0 Got 0     Should be     Sent 0 got 7
Sent 1 Got 0     Should be     Sent 1 got 7
Sent 2 Got 0     Should be     Sent 2 got 7
Sent 3 Got 7     Should be     Sent 3 got 7
Sent 4 Got 7     Should be     Sent 4 got 0
Sent 5 Got 7     Should be     Sent 5 got 0
Sent 6 Got 7     Should be     Sent 6 got 0
Sent 7 Got 7     Should be     Sent 7 got 0

Okay... with "<=3" at least 3 should be the same as 2...yes? And why do 0,1,2 get a return of 0 when they should get a return of 7?

You've found another bug. If you use Byte types instead, it works ok. The <= operator was bugged of Ubyte type. It's been fixed.
I guess relational operators need an intensive test :oops: The problem is I can't create tests (I currently run more than 200 test by hand on each compile). Tests need to be automatized. The problem is you can't check the tests results, since they must be run on a real Z80 or on a Z80 emulator. I'm currently developing one.

[britlion]

I think I have this listed as an issue in my summary post. Glad that I'm not the only one seeing issues with it, though. I was beginning to think I was going crazy