c++ - 来自std::string to unsigned char[], an error “non-constant-expression cannot be narrowed”

Question

没有错误消息:non-constant-expression cannot be narrowed from type 'std::basic_string<char, std::char_traits<char>, std::allocator<char> >::value_type' (aka 'char') to 'unsigned char' in initializer list [-Wc++11-narrowing]。
MSVC 和 GCC 将问题视为警告，但是 Clang 将其视为错误。
编码:

enum Encoding { ERR = -1, NONE, ASCII, UTF8, UTF16LE, UTF16BE, UTF32LE, UTF32BE };

Encoding CheckBOM(const std::string& data) {
    unsigned int DataSize = data.length();
    if (DataSize < 4)
        return ERR;
    // the following line is the cause of the problem
    unsigned char BOM[4] = { data[0], data[1], data[2], data[3] }; // <-- This line is what the error message points to

    if (BOM[0] == 0xef && BOM[1] == 0xbb && BOM[2] == 0xbf)
        return Encoding::UTF8;
    else if (BOM[0] == 0xff && BOM[1] == 0xfe && BOM[2] == 0x00 && BOM[3] == 0x00)
        return Encoding::UTF32LE;
    else if (BOM[0] == 0x00 && BOM[1] == 0x00 && BOM[2] == 0xfe && BOM[3] == 0xff)
        return Encoding::UTF32BE;
    else if (BOM[0] == 0xff && BOM[1] == 0xfe)
        return Encoding::UTF16LE;
    else if (BOM[0] == 0xfe && BOM[1] == 0xff)
        return Encoding::UTF16BE;
    return Encoding::NONE;
}

我曾尝试解决该问题，但由于不了解，所以无法解决。
我该如何解决？

Answer 1

由于std::string是basic_string<char>，因此您要为char分配带符号的unsigned char类型，从而缩小范围。
狭义的转换可以说是隐式转换，其中转换后的类型不能在所有情况下都由目标类型表示，这种情况就是将有符号转换为无符号。
由于该标准不允许在聚合初始化(即braced-init-list)中缩小转换范围，因此您必须自己“手动”完成操作。
dcl.init.aggr#4.2

[...] If that initializer is of the form assignment-expression or = assignment-expression and a narrowing conversion ([dcl.init.list]) is required to convert the expression, the program is ill-formed.[...]

dcl.init.list#7

A narrowing conversion is an implicit conversion

[...]

[可能的隐式转换列表，分析中不需要此列表]

[...]

[Note: As indicated above, such conversions are not allowed at the top level in list-initializations.— end note]

[Example:

[...]

unsigned char uc1 = {5}; // OK: no narrowing needed

unsigned char uc2 = {-1}; // error: narrows

unsigned int ui1 = {-1}; // error: narrows

[...]

为了避免错误/警告，您可以将其强制转换为 unsigned char:

unsigned char BOM[4] = { static_cast<unsigned char>(data[0]), static_cast<unsigned char>(data[1]), static_cast<unsigned char>(data[2]), static_cast<unsigned char>(data[3])};

Live demo
或者，作为 suggested by @RemyLebeau，您可以使用:

const unsigned char * pdata = reinterpret_cast<const unsigned char *>(data.c_str()); 
unsigned char BOM[4] = { pdata[0], pdata[1], pdata[2], pdata[3] }; 

或更简单:

unsigned char BOM[4]; 
std::memcpy(BOM, data.c_str(), 4);

这需要 #include<cstring>。