casting int to char using C++ style casting [dupli

问题:

This question already has answers here:

Regular cast vs. static_cast vs. dynamic_cast [duplicate] (8 answers)

Why use static_cast<int>(x) instead of (int)x? (9 answers)

Closed 6 years ago.

In traditional C you can do:

int i = 48;
char c = (char)i;
//Now c holds the value of 48. 
//(Of course if i > 255 then c will not hold the same value as i).  

Which of the c++ casting methods (static_cast, reinterpret_cast) is suited for getting this job done?

回答1:

You can implicitly convert between numerical types, even when that loses precision:

char c = i;

However, you might like to enable compiler warnings to avoid potentially lossy conversions like this. If you do, then use static_cast for the conversion.

Of the other casts:

• dynamic_cast only works for pointers or references to polymorphic class types;
• const_cast can't change types, only const or volatile qualifiers;
• reinterpret_cast is for special circumstances, converting between pointers or references and completely unrelated types. Specifically, it won't do numeric conversions.
• C-style and function-style casts do whatever combination of static_cast, const_cast and reinterpret_cast is needed to get the job done.

回答2:

You should use static_cast<char>(i) to cast the integer i to char.

reinterpret_cast should almost never be used, unless you want to cast one type into a fundamentally different type.

Also reinterpret_cast is machine dependent so safely using it requires complete understanding of the types as well as how the compiler implements the cast.

For more information about C++ casting see:

• When should static_cast, dynamic_cast, const_cast and reinterpret_cast be used?
• http://www.cplusplus.com/doc/tutorial/typecasting/.

回答3:

reinterpret_cast cannot be used for this conversion, the code will not compile. According to C++03 standard section 5.2.10-1:

Conversions that can be performed explicitly using reinterpret_cast are listed below. No other conversion can be performed explicitly using reinterpret_cast.

This conversion is not listed in that section. Even this is invalid:

long l = reinterpret_cast<long>(i)

static_cast is the one which has to be used here. See this and this SO questions.

回答4:

Using static cast would probably result in something like this:

// This does not prevent a possible type overflow
const char char_max = -1;

int i = 48;
char c = (i & char_max);

To prevent possible type overflow you could do this:

const char char_max = (char)(((unsigned char) char(-1)) / 2);

int i = 128;
char c = (i & char_max); // Would always result in positive signed values.

Where reinterpret_cast would probably just directly convert to char, without any cast safety. -> Never use reinterpret_cast if you can also use static_cast. If you're casting between classes, static_cast will also ensure, that the two types are matching (the object is a derivate of the cast type).

If your object a polymorphic type and you don't know which one it is, you should use dynamic_cast which will perform a type check at runtime and return nullptr if the types do not match.

IF you need const_cast you most likely did something wrong and should think about possible alternatives to fix const correctness in your code.