[Doxygen-users] Preprocessor templates

[byron.hawkins <by...@ha...>]

I'm trying to generate docs for a simple vector in C, which uses preprocessor
statements to simulate a C++ template. It's not possible to use C++ because
the vector is used in an in-process virtualization layer that is written in
C, and also it is required to use the heap in very restricted ways. The
vector's init() function looks like:

    bool
    VECTOR_NAME(,NAME_KEY,_init)(VECTOR_TYPE *vec, uint initial_capacity,
bool synch
        _IF_ELEMENT_POINTER(, void (*free_data_func)(ELEMENT_TYPE))
        _IF_SORTED(, int (*comparator)(ELEMENT_TYPE, COMPARISON_TYPE)));

All the tokens in capitals are defined by the client module above the
`#include`, so for example a client source file might have:

    #define NAME_KEY account_vector
    #define ELEMENT_TYPE account_t*
    #define COMPARISON_TYPE unsigned int
    #define SORTED 1
    #include "tvectorx.h"

This would generate a function:

    bool
    account_vector_init(account_vector_t *vec, unsigned int
initial_capacity, bool synch,
        void (*free_data_func)(account_t*),
        int (*comparator)(account_t*, unsigned int));

Not surprisingly, doxygen is not able to parse all the preprocessor stuff.
It comes out like this in the html:

    bool tvector_init (VECTOR_NAME(, NAME_KEY, _t)*vec, uint
initial_capacity, 
        bool synch, void(*free_data_func)(void *), 
        int(*comparator)(void *, VECTOR_COMPARISON_TYPE))

(Some default substitutions have been made.) Is there a way to just override
the doxygen function declaration with a plain string, so that doxygen
displays the string instead of what it would normally parse out of the code?
Thanks for your help.




--
View this message in context: http://doxygen.10944.n7.nabble.com/Preprocessor-templates-tp6540.html
Sent from the Doxygen - Users mailing list archive at Nabble.com.