Rule Definition
An anonymous namespace will be unique within each translation unit. Any declarations appearing in an anonymous namespace in a header will refer to different entities in each translation unit, which may not be consistent with developer expectations. Production quality C++ code uses header files to share code between translation units through #include directive. If an anonymous namespace is defined then each translation unit will define its own instance of the unnamed namespace. This may cause: Unexpected behavior, undefined behavior and may bloat the resulting executable

Remediation
Namespaces in header files should always be named.

Violation Code Sample

// Header.hpp
namespace // Non-compliant
{
extern int32_t x;
}

// File1.cpp
#include "Header.hpp"
namespace
{
int32_t x;
}
void fn_a ( void )
{
x = 24;
}
// File2.cpp
#include "Header.hpp"
namespace
{
int32_t x;
}
void fn_b ( void )
{
fn_a ( );
if ( x == 24 ) // This x will not have been initialized.
{
}
}

Standards Reference MISRA C++ 2008, 7-3-3: There shall be no unnamed namespaces in header files