Skip to end of metadata
Go to start of metadata

It is necessary to understand how macro replacement works in C, particularly in the context of concatenating tokens using the ## operator and converting macro parameters to strings using the # operator.

Concatenating Tokens

The ## preprocessing operator is used to merge two tokens into one while expanding macros, which is called token pasting or token concatenation. When a macro is expanded, the two tokens on either side of each ## operator are combined into a single token that replaces the ## and the two original tokens in the macro expansion [FSF 2005].

Token pasting is most useful when one or both of the tokens come from a macro argument. If either of the tokens next to a ## is a parameter name, it is replaced by its actual argument before ## executes. The actual argument is not macro expanded first.


Parameters are not replaced inside string constants, but the # preprocessing operator can be used instead. When a macro parameter is used with a leading #, the preprocessor replaces it with the literal text of the actual argument converted to a string constant [FSF 2005].

Noncompliant Code Example

The following definition for static_assert() from DCL03-C. Use a static assertion to test the value of a constant expression uses the JOIN() macro to concatenate the token assertion_failed_at_line_ with the value of __LINE__:

#define static_assert(e) \
  typedef char JOIN(assertion_failed_at_line_, __LINE__) \
    [(e) ? 1 : -1]

__LINE__ is a predefined macro name that expands to an integer constant representing the presumed line number of the current source line within the current source file. If the intention is to expand the __LINE__ macro, which is likely the case here, the following definition for JOIN() is noncompliant because the __LINE__ is not expanded, and the character array is subsequently named assertion_failed_at_line___LINE__:

#define JOIN(x, y) x ## y

 Compliant Solution

To get the macro to expand, a second level of indirection is required, as shown by this compliant solution:

#define JOIN(x, y) JOIN_AGAIN(x, y)
#define JOIN_AGAIN(x, y) x ## y

JOIN(x, y) calls JOIN_AGAIN(x, y) so that if x or y is a macro, it is expanded before the ## operator pastes them together.

Note also that macro parameters cannot be individually parenthesized when concatenating tokens using the ## operator, converting macro parameters to strings using the # operator, or concatenating adjacent string literals. This is an exception, PRE01-C-EX2, to PRE01-C. Use parentheses within macros around parameter names.

Noncompliant Code Example

This example is noncompliant if the programmer's intent is to expand the macro before stringification:

#define str(s) #s
#define foo 4


The macro invocation str(foo) expands to foo.

Compliant Solution

To stringify the result of expansion of a macro argument, two levels of macros must be used:

#define xstr(s) str(s)
#define str(s) #s
#define foo 4

The macro invocation xstr(foo) expands to 4 because s is stringified when it is used in str(), so it is not macro expanded first. However, s is an ordinary argument to xstr(), so it is completely macro expanded before xstr() is expanded. Consequently, by the time str() gets to its argument, it has already been macro expanded.

Risk Assessment




Remediation Cost









Automated Detection

Axivion Bauhaus Suite


Macro uses # operator
Macro uses ## operator
Helix QAC


C0341, C0342, C0801, C0802, C0803, C0811, C0872, C0880, C0881, C0884

LDRA tool suite

76 S, 125 S, 637 S

Enhanced Enforcement
PC-lint Plus



Assistance provided: reports any use of pasting or stringizing operators in a macro definition


0341, 0342, 0801, 0802, 803, 0811, 0872,

0880, 0881, 0884

Related Vulnerabilities

Search for vulnerabilities resulting from the violation of this rule on the CERT website.

Related Guidelines


[FSF 2005]Section 3.4, "Stringification"
Section 3.5, "Concatenation"
[Saks 2008]


  1. Compliant Solution

    To stringify the result of expansion of a macro argument, you must use two levels of macros:

    #define xstr(s) str(s)
    #define str(s) #s
    #define foo 4
    Isn't it wrong to use #define foo 4  
    I think it should be #define foo (4) to comply with standard Use parentheses within macros around parameter names
    1. Unfortunately, that would wind up printing (4) instead of 4 when using the xstr macro to expand foo.

    2. Writing:

          #define FOO 22

      is fine because the pre-processor simply replaces a token with another one and the semantics is the same.  What is bad is if you do:

         #define FOO -22

      In that case you need to use parentheses to avoid problems, as in:

        #define FOO (-22)

      I think Use parentheses within macros around parameter names should be updated to describe this.

  2. Wouldn't it make sense to use an array size of zero instead of one if the assertion does not fail?

    #define static_assert(e) \
      typedef char JOIN(assertion_failed_at_line_, __LINE__) \
        [(e) ? 0 : -1]

    Sure, just an example but adding a lot of that static_assert will (if unused variables are not optimized out) consume some stacks size using 1 and shouldn't using 0.

    Or is this assumption wrong?

    1. Zero-sized arrays are not allowed by ISO C, so that wouldn't work portably. Thankfully, there's not a storage issue here to be solved; typedefs do not define a new variable, only a type alias, so there is no runtime overhead from their use. See the output (even at -O0) here as an example: and note how there's space reserved for c  but not foo  or bar .