指向 POD 结构的 C/C++ 指针也指向第一个结构成员

Question

Can I assume that a C/C++ struct pointer will always point to the first member?
Example 1:

typedef struct {
 unsigned char  array_a[2];
 unsigned char  array_b[5];
}test;
//..
test var;
//..

In the above example will &var always point to array_a? Also in the above example is it possible to cast the pointer to an unsigned char pointer and access each byte separately?
Example 2:

function((unsigned char *)&var,sizeof(test));
//...
//...
void function(unsigned char *array, int len){
 int i;
 for( i=0; i<len; i++){
    array[i]++;
 }
}

Will that work correctly?

Note: I know that chars are byte aligned in a struct therefore I assume the size of the above struct is 7 bytes.

Answer 1

For C structs, yes, you can rely on it. This is how almost all "object orientated"-style APIs work in C (such as GObject and GTK).

For C++, you can rely on it only for "plain old data" (POD) types, which are guaranteed to be laid out in memory the same way as C structs. Exactly what constitutes a POD type is a little complicated and has changed between C++03 and C++11, but the crux of it is that if your type has any virtual functions then it's not a POD.

(In C++11 you can use std::is_pod to test at compile-time whether a struct is a POD type.)

EDIT: This tells you what constitutes a POD type in C++: http://en.cppreference.com/w/cpp/concept/PODType

EDIT2: Actually, in C++11, it doesn't need to be a POD, just "standard layout", which is a lightly weaker condition. Quoth section 9.2 [class.mem] paragraph 20 of the standard:

A pointer to a standard-layout struct object, suitably converted using a reinterpret_cast, points to its initial member (or if that member is a bit-field, then to the unit in which it resides) and vice versa. [ Note: There might therefore be unnamed padding within a standard-layout struct object, but not at its beginning, as necessary to achieve appropriate alignment. — end note ]

Answer 2

From the C99 standard section 6.7.2.1 bullet point 13:

Within a structure object, the non-bit-field members and the units in which bit-fields reside have addresses that increase in the order in which they are declared. A pointer to a structure object, suitably converted, points to its initial member (or if that member is a bit-field, then to the unit in which it resides), and vice versa. There may be unnamed padding within a structure object, but not at its beginning.

The answer to your question is therefore yes.

Reference (see page 103)

Answer 3

The compiler is free to add padding and reorganize the struct how it sees fit. Especially in C++ you can add (virtual) functions and then chances are that the virtual table is hidden before that. But of course that are implementation details. For C this assumption is valid.

Answer 4

For C, it's largely implementation-specific, but in practice the rule (in the absence of #pragma pack or something likewise) is:

• Struct members are stored in the order they are declared. (This is required by the C99 standard, as mentioned here earlier.)
• If necessary, padding is added before each struct member, to ensure correct alignment.

So given a struct like

struct test{
char ch;
int i;
}

will have ch at offset 0, then a padding byte to align, i at offset 2 and then at the end, padding bytes are added to make the struct size a multiple of 8 bytes.(on a 64-bit machine, 4 byte alignment may be permitted in 32 bit machines)

So at least in this case, for C, I think you can assume that the struct pointer will point to the first array.