Norcroft c 5.88 generates infinite loop

Hello all,
A program of mine went into a loop and I was suprised to find that cc 5.88 had
generated the equivalent of:

xxxx B xxxx

Closer inspection showed that it was a bug in my program but I am wondering if
Norcroft have put out a warning message about the code. I would be interested
to know what other people think.

The actual code went something like this:

extern int aaa(void);

while (aaa != 0);

The bug is that it should have said:

while (aaa() != 0);

I can see what is happening here. It has taken ‘aaa’ as a constant that
is not equal to zero and optimised the expression. The condition will
always be true so it has generated:

xxxx B xxxx

I would just put it down a mistake on my part if ‘aaa’ was variable
containing the address of the function but it is not. It is effectively
a constant. I think it would be helpful if Norcroft put out a warning
message if it found a reference to a function in a strange context such
as in the conditional part of a ‘while’ statement. What do other people
think? Is it just my dodgy programming or do you think that detecting
this sort of this and putting out warning messages would be helpful?

Dave Daniels

For comparison, clang doesn’t warn about the infinite loop, but it does warn about the comparison always being true:

test.c:8:8: warning: comparison of function 'aaa' not equal to a null pointer is always true [-Wtautological-pointer-compare]
        while(aaa != 0);
              ^~~    ~
test.c:8:8: note: prefix with the address-of operator to silence this warning
        while(aaa != 0);
              ^
              &
test.c:8:8: note: suffix with parentheses to turn this into a function call
        while(aaa != 0);
              ^
                 ()

Applying either of the two suggested changes will remove the warning, with no other warnings being generated (this is with -Wall). So it looks like the warning is only produced because the compiler thinks you’ve made a typo (which is exactly what the problem was, in your case).

I’m not sure I’ve ever seen a C compiler warn about infinite loops.

Hello Jeffrey,
Thanks for your reply. I was thinking about reporting that the name of the function is being used in a strange, possibly incorrect context rather than reporting on the code generated. Norcroft does produce at least one warning already when it detects ‘=’ in a conditional context but I imagine that this one is a far more common typing error than my one but there is a precedence for the compiler reporting possible errors. But then, where do you stop? I have been dabbling in C since the mid 1980s (Hisoft C, anyone?). I remember someone once saying that you can learn C in an afternoon and then spend the rest of your life mastering it. I know what he means!

Dave Daniels