Function won't return anything other than 0.

[object01]

I've been wrestling with what seems like a compiler bug. I'm hoping somebody can point out something I've missed.

The following code fragment from my program generates the listing that follows. There are two things I don't understand: first, the instruction at 0xA3A, GOTO 23B, actually goes to 0xA3B. Second, the only return statement in any of this is the RETLW 00, which according to the PIC16F88 data sheet will return 0. Sure enough, the function this code fragment comes from never returns anything but 0.

Notice the two commented return statements. If I uncomment those statements and comment out the return (matchFound), I get the same problem. The program eventually flows to the RETLW 00.

I'm using PCM v3.189.

--
Jeff S.

[rnielsen]

I thought that a return() (or return;) would break out of the function and not execute the remaining code of the function. If you un-comment the first two return's then the third one would never be executed. Plus, I think if you want to return a value you need to have return(value) stated. Just a simple 'return true;' or 'return false;' will simply return with no value being returned. You might want to change it to return(true); and return(false);. But, then again, if you do, the third return(matchFound) would not be executed.

Ronald

[William H. Conley III]

Try..........



if (!matchFound)
{
fprintf(DEBUGGER, "\nDid not match ");
fprintf(DEBUGGER, ATINFO.response);
fprintf(DEBUGGER, "\n");
return (FALSE);
}
else
{
fprintf(DEBUGGER, "\nMatched ");
fprintf(DEBUGGER, ATINFO.response);
fprintf(DEBUGGER, "\n");
return (TRUE);
}
return (matchFound);

[object01]

Including the commented "return TRUE" and "return FALSE" statements results in the same problem. In fact, they cause the program to eventually flow to the same RETLW 00 statement. And parentheses around any of the return statements' parameters don't seem to affect it either. The assembly listing just keeps coming down to that RETLW 00 statement.

--
Jeff S.

[Guest]

It has to be a bug? Your code is nothing I haven't used my self

[Guest]

>> first, the instruction at 0xA3A, GOTO 23B, actually goes to 0xA3B.

This is the effect of ... 0A36: BSF 03.6

>> Second, the only return statement in any of this is the RETLW 00,

The return value is ... 0A39: MOVWF 78 ... CCS use RAM 0x78 for return value instead of the W.

read the SYM file:

078 @SCRATCH
078 _RETURN_
079 @SCRATCH

and read the LST file after the function call, there will be a line ... MOVF 78,W ... right after the CALL xxx.

[Ed Arnold]

I have had the same problem with my code. I ended up scraping the code. I am using version 3.148? Keep us posted with any new finding you may have.