Skip to content

Variables I and Arithmetic Types

TLDR 见 Summary / TLDR

Variable declaration

Type

  • Type can not be changed
  • Known even when the program is not run
    • known at compile-time
    • C is statically-typed <=> C has a static type system.

Arithmetic types

Refer this

Integer

  • signed or unsigned
  • int is optional (short int = short, unsigned int = unsigned)
  • unsigned needs to be written
    • Types without the keyword unsigned are signed by default
type width (at least) width (usually)
short 16 bits 16 bits
int 16 bits 32 bits
long 32 bits 32 or 64 bits
long long 64 bits 64 bits

  1. Minimum Possible Width (最小可能宽度)：这指的是该类型在任何平台上的最小位宽度，C标准规定数据类型的最小宽度。例如，short类型被规定至少需要16位。这确保了程序在不同平台上的可移植性。

  2. Typical Width (典型宽度)：这指的是该类型在大多数常用平台上的典型位宽度。因为硬件和操作系统之间的差异，对于某些类型，尤其是int和long类型，它们的宽度可能在不同系统上有所不同。例如，int类型在大多数现代系统上通常是32位，但也有一些旧系统使用16位。

  • A signed type has the same width as its unsigned counterpart.

Implementation-defined behaviors 实现定义行为

The standard states that the exact width of the integer types is implementation-defined.

  • Implementation: The compiler and the standard library.
  • An implementation-defined behavior depends on the compiler and the standard library, and is often also related to the hosted environment (e.g. the operating system).

Real floating types

C has three types for representing real floating-point values:

  • float : single precision. Matches IEEE754 binary32 format if supported.
  • double : double precision. Matches IEEE754 binary64 format if supported.
  • long double : extended precision. A floating-point type whose precision and range are at least as good as those of double .

Range of Values

Do not use floating-point types for integer arithmetic!

Addition about limitations of floating types

1. 不能乱用

double midpoint(long long a, long long b) {
return (a + b) / 2.0;
}
// 9007199254740993 9007199254740993

2. 浮点误差

判断一个浮点数是否等于零, 必须使用 \(|x|<\epsilon\), 其中 \(\epsilon\) 可以是 \(10^{-8}\) 这样的一个小数。判断两个浮点数是否相等, 必须使用 \(|a-b|<\epsilon\) 。具体参考 IEEE 754

以下的例子输出：0.19999999999999998

void test_double(void) {
double a = 1;
double b = a / 3;
return b * 3;
}

scanf / printf

Reference

type format specifier
short %hd
int %d
long %ld
long long %lld
type format specifier
unsigned short %hu
unsigned %u
unsigned long %lu
unsigned long long %llu
type format specifier
%f float
%lf double
%Lf long double

Charater Type

  • Size: 1 byte (any char type is smaller than short, int, ...)
    • signed char: \([-128, 127]\)
    • unsigned char: \([0, 255]\)
  • Escape Charater
  • signed / unsigned: Implementation-defined
  • %c
  • Questions on non-ASCII characters

Boolean Type

  • Since C99: 1, 0 => true false
  • #include <stdbool.h>

Summary / TLDR

../res/Pasted image 20240331175601.png - 1 byte = 8 bit - sizeof(T) 可以获得 T 所占的字节数量 - printf("%zu\n", sizeof(double));