Computing the absolute value of an integer

[Shawn H. Oesterle]

In article <8977 at hydra.gatech.EDU> dsrekml at prism.gatech.EDU (Mike Mitten) writes:
>In article <1990May4.121950.22726 at agate.berkeley.edu> c60c-3cf at e260-3c.berkeley.edu (Dan Kogai) writes:
>>What's wrong with using a macro like the following
>>
>>#define abs(x) (((x) >= 0) ? (x) : -(x))

>Won't both of these macros blow up if x is a function such as:
>
>	y = abs(foo(bar));
>
>or
>
>	y = _abs(foo(bar));
>

That's right.  That is one thing which I dislike about macros in C.  For
example, take the following function which finds the height of a binary
tree.   
 
 
#define max(a,b)    (((a) > (b)) ? (a) : (b))

/* This function returns the height in O(n) time (n is the number of nodes) */

int TreeHeight(const TREE * tree)  {
     if(tree != NULL)  {
          int  x = TreeHeight(tree->left),
               y = TreeHeight(tree->right);
 
          return(1 + max(x,y));
     }
     else  {
          return(-1);
     }
}
 

/* because the 'max' macro has functions as its arguments, TreeHeight() is
   changed from a linear time complexity function to an exponential time
   complexity function! */

int TreeHeight(const TREE * tree)  {
     if(tree != NULL)  {
          return(1 + max(TreeHeight(tree->left), TreeHeight(tree->right)));
     }
     else  {
          return(-1);
     }
}

BTW, C++ resolves this problem with the 'inline' keyword for function
declarations.

-- 
Shawn Oesterle { oesterle at wpi.wpi.edu } antherea

"...I tend to think that there are only two kinds of people in 
the world anyway: those that oversimplify and those that don't."  -Jim Northrup