Constants

Character constants

a character constant can take one of four forms:
- a single character enclosed in a pair of single quotes, like 'a' or '$'
- a backslash character followed by one to three octal digits (any digit from 0 to 7), like '\7' or '\007'
- a backslash character followed by an 'x' (but not 'X') and one or two hexadecimal digits (any digit from 0 to 9 as well as any letter from 'a' to 'f' or from 'A' to 'F'), like '\xa0' or '\03F'
- a backslash character followed by one of eleven special characters, which are interpreted like:
```
	\a	alert (bell) character
	\b	backspace
	\f	formfeed
	\n	newline
	\r	carriage return
	\t	horizontal tab
	\v	vertical tab
	\\	backslash
	\?	question mark
	\'	single quote
	\"	double quote
```
ANSI C also allows Lcharacter-constant to define a wide character, using a group of characters and escape codes to form a single multibyte character for alphabets (Japanese, Chinese, etc) with more than 256 symbols.
some implementations may also allow multi-character constants, where a constant like 'PICT' is packed into a single 32 bit word
- the byte order of a packed multi-character constant is machine-dependent and thus non-portable

Integer constants

Integer constants can be one of the different bases:
- a number starting with a digit from 1 to 9 followed only by digits from 0 to 9 is a decimal (base 10) constant, like 1 or 54321
- if a number begins with '0x' or '0X' and is followed by digits from 0 to 9 and letters from a to f or from A to F, it is interpreted as a hexadecimal (base 16) constant, like 0xbeef or 0XDEAD
- if a number begins with 0 and is followed by digits from 0 to 7, it is interpreted as an octal (base 8) constant, like 07 (which is the same as 7) or 0111 (which is NOT the same as 111)
integer constants in any of the three bases may be followed by an 'l' or 'L' to indicate a long constant, like 17l or 0xFfL
in ANSI C, an integer constant may be followed by:
- a 'u' or 'U' to indicate an unsigned constant, like 33u
- a 'ul' or 'UL' to indicate an unsigned long constant, like 32767ul
an integer constant takes on the minimum type necessary to be accurately represented, (possibly) starting as int and increasing from long to unsigned long
- octal and hexadecimal constants, and (in ANSI C) constants ending with a 'u' or 'U' may also be unsigned (which can hold larger values than int but may have fewer bytes than long)
in ANSI C, 0 is always unsigned (perhaps to reduce ambiguity on machines with one's complement arithmetic)

as an example, if an int is 16 bits wide and a long is 32 bits wide, the following ranges would apply:

Range			Decimal		Octal/Hexadecimal
0			unsigned	unsigned
1-32767			int		int
32768-65535		long		unsigned
65536-2147483647	long		long
2147483648-429497295	unsigned long	unsigned long
429497296-?		???		???

Floating-point constants

a floating-point constant can take one of two forms:
- a group of digits from 0 to 9 which includes a single period, like 3.141593, 0. or .1
- a group of digits from 0 to 9, possibly containing a period, followed by 'e' or 'E' and an exponent, like 1.0e3, 2.593e+7 or 3141593e-6
in ANSI C, a floating-point constant may be followed by:
- an 'f' or 'F' to indicate a float constant, like 35.79f
  - Note that 123f is an error, because 123 is not a floating-point constant
- a 'l' or 'L' to indicate a long double constant, like 35.79l
in the absence of a 'f' or 'l' qualifier, floating-point constants are assumed to be of datatype double, which should be the most efficient floating-point datatype

Previous, Next, Index