When an operator has operands of different types, they are converted to a common
type according to a small number of rules. In general, the only automatic
conversions are those that convert a ``narrower'' operand into a ``wider'' one
without losing information, such as converting an integer into floating point in
an expression like f + i. Expressions that don't make sense, like using
a float as a subscript, are disallowed. Expressions that might lose
information, like assigning a longer integer type to a shorter, or a
floatingpoint type to an integer, may draw a warning, but they are not illegal.
A char is just a small integer, so chars may be freely used
in arithmetic expressions. This permits considerable flexibility in certain
kinds of character transformations. One is exemplified by this naive
implementation of the function atoi, which converts a string of digits
into its numeric equivalent.
/* atoi: convert s to integer */
int atoi(char s[])
{
int i, n;
n = 0;
for (i = 0; s[i] >= '0' && s[i] <= '9'; ++i)
n = 10 * n + (s[i]  '0');
return n;
}
the expression
s[i]  '0'
gives the numeric value of the character stored in s[i], because the
values of '0', '1', etc., form a contiguous increasing
sequence.
Another example of char to int conversion is the function lower,
which maps a single character to lower case for the ASCII character set.
If the character is not an upper case letter, lower returns it
unchanged.
/* lower: convert c to lower case; ASCII only */
int lower(int c)
{
if (c >= 'A' && c <= 'Z')
return c + 'a'  'A';
else
return c;
}
This works for ASCII because corresponding upper case and lower case letters are
a fixed distance apart as numeric values and each alphabet is contiguous 
there is nothing but letters between A and Z. This latter
observation is not true of the EBCDIC character set, however, so this code would
convert more than just letters in EBCDIC.
The standard header <ctype.h>, defines a family of functions
that provide tests and conversions that are independent of character set. For
example, the function tolower is a portable replacement for the
function lower shown above. Similarly, the test
c >= '0' && c <= '9'
can be replaced by
isdigit(c)
We will use the <ctype.h> functions from now on.
There is one subtle point about the conversion of characters to integers. The
language does not specify whether variables of type char are signed or
unsigned quantities. When a char is converted to an int, can
it ever produce a negative integer? The answer varies from machine to machine,
reflecting differences in architecture. On some machines a char whose
leftmost bit is 1 will be converted to a negative integer (``sign extension'').
On others, a char is promoted to an int by adding zeros at the left
end, and thus is always positive.
The definition of C guarantees that any character in the machine's standard
printing character set will never be negative, so these characters will always
be positive quantities in expressions. But arbitrary bit patterns stored in
character variables may appear to be negative on some machines, yet positive on
others. For portability, specify signed or unsigned if
noncharacter data is to be stored in char variables.
Relational expressions like i > j and logical expressions
connected by && and  are defined to have value 1 if
true, and 0 if false. Thus the assignment
d = c >= '0' && c <= '9'
sets d to 1 if c is a digit, and 0 if not. However, functions
like isdigit may return any nonzero value for true. In the test part
of if, while, for, etc., ``true'' just means
``nonzero'', so this makes no difference.
Implicit arithmetic conversions work much as expected. In general, if an
operator like + or * that takes two operands (a binary
operator) has operands of different types, the ``lower'' type is promoted
to the ``higher'' type before the operation proceeds. The result is of the
integer type. If there are no unsigned operands, however, the following
informal set of rules will suffice:
 If either operand is long double, convert the other to long
double.
 Otherwise, if either operand is double, convert the other to double.
 Otherwise, if either operand is float, convert the other to float.
 Otherwise, convert char and short to int.
 Then, if either operand is long, convert the other to long.
Notice that floats in an expression are not automatically converted to double;
this is a change from the original definition. In general, mathematical
functions like those in <math.h> will use double precision. The
main reason for using float is to save storage in large arrays, or,
less often, to save time on machines where doubleprecision arithmetic is
particularly expensive.
Conversion rules are more complicated when unsigned operands are
involved. The problem is that comparisons between signed and unsigned values are
machinedependent, because they depend on the sizes of the various integer
types. For example, suppose that int is 16 bits and long is 32
bits. Then 1L < 1U, because 1U, which is an unsigned
int, is promoted to a signed long. But 1L > 1UL
because 1L is promoted to unsigned long and thus appears to
be a large positive number.
Conversions take place across assignments; the value of the right side is
converted to the type of the left, which is the type of the result.
A character is converted to an integer, either by sign extension or not, as
described above.
Longer integers are converted to shorter ones or to chars by
dropping the excess highorder bits. Thus in
int i;
char c;
i = c;
c = i;
the value of c is unchanged. This is true whether or not sign extension
is involved. Reversing the order of assignments might lose information, however.
If x is float and i is int, then x =
i and i = x both cause conversions; float to int
causes truncation of any fractional part. When a double is converted to
float, whether the value is rounded or truncated is implementation
dependent.
Since an argument of a function call is an expression, type conversion also
takes place when arguments are passed to functions. In the absence of a function
prototype, char and short become int, and float
becomes double. This is why we have declared function arguments to be int
and double even when the function is called with char and float.
Finally, explicit type conversions can be forced (``coerced'') in any
expression, with a unary operator called a cast. In the construction
(type name) expression
the expression is converted to the named type by the conversion
rules above. The precise meaning of a cast is as if the expression were
assigned to a variable of the specified type, which is then used in place of the
whole construction. For example, the library routine sqrt expects a double
argument, and will produce nonsense if inadvertently handled something else. (sqrt
is declared in <math.h>.) So if n is an integer, we can
use
sqrt((double) n)
to convert the value of n to double before passing it to sqrt.
Note that the cast produces the value of n in the proper type;
n itself is not altered. The cast operator has the same high precedence
as other unary operators, as summarized in the table at the end of this chapter.
If arguments are declared by a function prototype, as the normally should be,
the declaration causes automatic coercion of any arguments when the function is
called. Thus, given a function prototype for sqrt:
double sqrt(double)
the call
root2 = sqrt(2)
coerces the integer 2 into the double value 2.0
without any need for a cast.
The standard library includes a portable implementation of a pseudorandom
number generator and a function for initializing the seed; the former
illustrates a cast:
unsigned long int next = 1;
/* rand: return pseudorandom integer on 0..32767 */
int rand(void)
{
next = next * 1103515245 + 12345;
return (unsigned int)(next/65536) % 32768;
}
/* srand: set seed for rand() */
void srand(unsigned int seed)
{
next = seed;
}
