# and # # in a C + + macro

The function of the # in a macro is to string the macro arguments that follow it (stringizing operator), simply by adding a double quote to the left and right of the macro variable it refers to.

The following two statements are equivalent, as defined #define STRING (x) #x .

char *pchar = "Hello";

Char *pchar = STRING (hello);

And a #@ is a single quote (Charizing Operator)

#define MAKECHAR (x) #@x

char ch = makechar (b); with char ch = ' B '; equivalence.

But there are small issues to note that the macro encounters # or # #时就不会再展开宏中嵌套的宏了. What's the point? For example, using Char *pchar =string (__file__), although __file__ itself is a macro, the compiler does not expand it, so PChar will point to "__file__" instead of the source file name of the form "D:\XXX.cpp" you want. Therefore, to add an intermediate conversion macro, first parse the __file__ into a "D:\XXX.cpp" string.

Define two macros as shown below:

#define _STRING (x) #x

#define STRING (x) _string (x)

Call the following statement again to output the source file path with ""

char* PChar = STRING (__file__);

printf ("%s%s\n", PChar, __file__);

You can compare the difference between the next string (__file__) and the __file__ , with double quotes before it, and no double quotes in the latter one.

Again, the next # #的功能, it can be spliced with symbols (token-pasting operator).

There is an example on MSDN:

#define PASTER (N) printf ("token" #n "=%d\n", token# #n)

int token9 = 100;

Call Paster (9) Again, and token# #n直接合并变成了token9 after the macro expands. The entire statement becomes a

printf ("token" "9" "=%d", token9);

In the C language, two concatenated double quotes in the string are automatically ignored, and the above sentence is equivalent to

printf ("Token9 =%d", token9);.

i.e. Output TOKEN9 = 100

With the basics above, we'll look at example 1.

#define WIDEN2 (x) L # # X

#define Widen (x) WIDEN2 (x)

#define __WFILE__ Widen (__file__)

wchar_t *pwsz = __wfile__;

The L in the first macro is the conversion of an ANSI string into a Unicode string. such as: wchar_t *pstr = L "Hello";

Then see wchar_t *pwsz = __wfile__;

__WFILE__ is first expanded into widen (__file__), then expanded into WIDEN2 ("__file__-represented string"), then stitched into L "__file__-represented string" i.e. L "D:\XXX.cpp" The Unicode string is obtained and the string address is assigned to the PWSZ pointer.

In VC _t (), TEXT () is also used in this technique.

In the Tchar.h header file, you can find:

#define _T (x) __t (x)

#define __T (x) L # # X

In the WinNT.h header file, you can find

#define TEXT (quote) __text (quote)//R_winnt

#define __text (quote) l# #quote//R_winnt

So it's easy to understand why the third statement will go wrong error C2065: ' Lsztext ': undeclared identifier

wprintf (Text ("%s%s\n"), _t ("Hello"), text ("Hello"));

Char sztext[] = "Hello";

wprintf (Text ("%s%s\n"), _t (Sztext), text (Sztext));

When you define "Hello" as a macro, you can run it correctly.

#define SZTEXT "Hello"

wprintf (Text ("%s%s\n"), _t (Sztext), text (Sztext));

# and # # in a C + + macro