Sizeof calculation space size summary

About sizeof Summary of
　　1、sizeof Use form of ：sizeof(var_name) perhaps sizeof var_name
　　2、 Variable names can be enclosed without parentheses ：sizeof a
　　3、 Data types must be enclosed in parentheses ：sizeof(int)
　　4、sizeof Expressions can be evaluated ：sizeof(2+3.14), the truth is that sizeof(double)
　　5、sizeof Function calls can be evaluated , It is actually evaluating the return value type
　　6、 The following situations cannot be used sizeof Is evaluated
　　 1) Function names cannot be evaluated
　　 2) Cannot evaluate the type of an indefinite return value , Such as void
　 3) Bit field members cannot use sizeof evaluation
　　7、sizeof The result is size_t, It is defined as unsigned int type . This type ensures that the byte size of the largest object created by the display can be accommodated .
　　 commonly , stay 32 Bit compiler ：

   sizeof(int):4
   sizeof(short):2
   sizeof(long):4
　 sizeof(long long):8
　 sizeof(float):4
　 sizeof(double):8
　 sizeof(char):1
   sizeof(p):4,(p For the pointer )

These values vary from system to system .
　　8、sizeof And strlen() Compare , When calculating a character array ,sizeof contain ’\0’,strlen() It doesn't contain ’\0’
　　9、32 Of the pointer under the bit system sizeof yes 4 byte ,64 Next is 8 byte
　　10、 Reference to sizeof, In fact, it is for the object to which it is bound sizeof
　　11、 about C character string , It needs to be remembered C/C++ One Chinese character takes up two bytes (Linux Next 3 Bytes ), And there is an empty character at the end of the string .
　　12、 An array sizeof You can calculate the size of the array
　　13、 Array as a parameter , The name of the array sizeof yes 4 Bytes
　　14、32bit In the system ,int **a[3][4];sizeof(a)=48
　　15、 About struct Space computing for
　　struct The space calculation of is complex , Generally, two rules are followed ：
　　 1) stay windows32 In the environment , The overall space is the member that occupies the largest space ( The type of ) An integer multiple of the occupied bytes , stay Linux+gcc In the environment , If the number of bytes occupied by the maximum member type exceeds 4, Such as double yes 8, Then the whole is empty 　　 Between is 4 An integer multiple of .
　　2) stay windows32 In the environment , Data vs. its principles —— The sequence of memory structure members , When it comes to the member variable , The size of the space placed in front of it must be an integer multiple of the size of the member type , If not enough 　　 Then make up , And so on . But in Linux+gcc In the environment , If the member type exceeds 4, Such as double yes 8, The size of the space already placed in front is 4 An integer multiple of , If it is not enough, it will be completed .
　　
　　win32 In the environment ： When it's placed , The train of thought is ：
1、 First place the first member at the address 0 The place of
2、 Then find the starting position of the second member according to its type size , The starting position of placement must be an integer multiple of the type size , By analogy .
3、 Last , Then calculate whether the overall space is an integer multiple of the maximum type size .
　　 And in the Linux+gcc In the environment : The idea of placement is
1、 First place the first member at the address 0 The place of
2、 If one unit cannot be installed , Then it is placed at the starting position of a unit address .
3、 Last , If the maximum type band exceeds 4, Then calculate whether the whole space is 4 Integer multiple .

16、 Space calculation of structures containing structures ：
1) stay windows32 In the environment , The whole space is between the child structure and the parent structure , The most occupied member ( The type of ) An integer multiple of the occupied bytes , stay Linux+gcc In the environment , If the number of bytes occupied by the maximum member type exceeds 4, Such as 　　double yes 8, Then the overall space is 4 An integer multiple of .
2) stay windows32 In the environment , Data vs. its principles —— The memory of the parent structure is arranged in the order of the structure members , When ranking to substructure members , The space placed in front of it must be the largest in the substructure members 　　 An integer multiple of the size of a large type , If it is not enough, it will be completed , One analogy .
　　win32 In the environment ： When it's placed , The train of thought is ：
1、 First place the first member at the address 0 The place of
2、 Then find the starting position of the second member according to its type size , The starting position of placement must be an integer multiple of the type size , If it is a substructure , Then the starting position of the placement is the maximum type size in the substructure 　　 Integer multiple , By analogy .
3、 Last , Then calculate whether the overall space is an integer multiple of the maximum type size in the parent-child structure .
　　Linux+gcc Save memory , Type size exceeds 4, The starting position should be 4 An integer multiple of .
　　17、 Space calculation of structures containing arrays ：
　　 In the structure , Arrays are placed one by one according to a single variable , Not as a whole , Therefore, the alignment of structures containing arrays is quite special .
　　 Such as ：
　　struct S1
　　{
　　 char a[8];
　　 int b;
　　};
　　sizeof(S1)=12, instead of 16.
　　
　　18、 Calculation of structures containing potential fields
　　 Bit field members cannot be deleted alone sizeof value , But structures containing bit fields can .
　　 There are the following rules for using bit fields ：
　　 1) If the types of adjacent bit fields are the same , And the sum of its bit width is less than that of type sizeof size , Then the following field will be immediately adjacent to the storage of the previous field , Until you know you can't hold it .
　　 stay vs2010 Next , One int a:4, If the following is not a bit field , Then occupy 4 Bytes , That is, the size of its type . And in the Dec-C++ And gcc Next , Whether it's a bit field or not , The number of bytes occupied shall be subject to the actual number of bytes occupied , namely int 　　a:4, If the following is not a bit field , Only one byte .
　　2) If the types of adjacent bit fields are the same , But the sum of its bit width is greater than that of its type sizeof size , The following fields will start with the new storage unit , Its offset is an integral multiple of the size of its type .
　　3) If the types of adjacent bit fields are different , Different compilers have different implementations ,VC6 No compression , and Dec-C++ And gcc Use compression .
　　4) If there are non bit field fields interspersed between bit field fields , No compression .
　　5) The total size of the whole structure is an integer multiple of the size of the fastest basic type member .

19、 Use "#pragma pack" Time structure space calculation
　　 Generally, you can change the default conditions by the following methods ：
　　 Using pseudo instructions #pragma pack(n), The compiler will follow n Byte alignment ;
　　 Using pseudo instructions #pragma pack(n), Custom byte alignment will be canceled .
　　 The above two instructions should usually be used together .
　　#pragma pack(n),n Is the number of byte alignments , Its value is 1、2、4、8、16 etc. , The default is 8, If this value is greater than the sizeof Small value , The offset of the member should be based on this , That is, the offset of the structure member should be 　　 Take the minimum of the two . The formula is as follows ：
　　offsetof(item)=min(n,sizeof(item))

20、union Space computing for
　　 Of the consortium sizeof Value is all members sizeof The maximum of .

21、 Enumerative spatial computation
　　enum It just defines a set of constants , There's no “ Elements ”, Enumeration is used as int Type stored , Therefore, the sizeof The value is 4.