Printable View
Hello,
I got a third party prototypes declaration as follow:
Code:
typedef struct mydata{
DWORD Length; //length of data field in bytes
BYTE Data[1]; // the data itself, variable length
}MYDATA
My question is why the Data is an array of 1???, if it is supposed to be of variable length, of value Length., as stated in the comments??
Because 1 is the smallest dimension for the array that allocates space for it. The way these things work is that the struct is declared with an array of size 1 (as the last member). When the memory for the struct is allocated it can be bigger than the structs declared size allowing access to further array elements.
eg. MYDATA *p = malloc(sizeof(MYDATA) + 100;
now the array indexes could range from 0 to 100.
The Length field is used to convey the number of elements that are accesible in the array.
the following achieve the same purpose, why not use the following style instead?
Code:
typedef struct mydata{
DWORD Length; //length of data field in bytes
BYTE * Data; // pointer to data
}MYDATA
Search google for "struct hack". You will find many hits and lots of information on why this strange way of setting up a structure is done.
Regards,
Paul McKenzie
May be just because BYTE* pointer occupies sizeof(WORD) bytes but BYTE occupies only 1. Further, due to different word sizes, sizeof structure does not remain constant - for compatibility between platforms, this may be imposed.
Also changing the type from BYTE to BYTE* moves the actual data out of the struct.
Quote:
Originally posted by Ajay Vijay
BYTE* pointer occupies sizeof(WORD) bytes
Code:sizeof( WORD ) = 2 bytes
sizeof( BYTE* /* or any other pointer on Win32 */ ) = 4 bytes
sizeof( BYTE** /* or any other pointer on Win32 */ ) = 4 bytes
Sorry, but I meant with word size as per the platform and not the WORD datatype. In other words what 'int' occupies (Generally) on various OSs.