Using the C Math Library

How mlfFeval() Works

mlfFeval() uses a built-in table to find out how to execute a particular function. The built-in table provides mlfFeval() with two pieces of information: a pointer that points to the function to be executed and a pointer to what's called a "thunk function."

As shipped, mlfFeval()'s built-in table contains each function in the MATLAB C Math Library. If you want mlfFeval() to know how to execute a function that you've written, you must extend the built-in table by creating a local function table that identifies your function for mlfFeval().

It's the thunk function, however, that actually knows how to execute your function. In Example - Passing Functions As Arguments (ex4.c), the thunk function, _lorenz_thunk_fcn_, executes lorenz(). A thunk function's actions are solely determined by the number of input and output arguments to the function it is calling. Therefore, any functions that have the same number of input and output arguments can share the same thunk function. For example, if you wrote three functions that each take two inputs and produce three outputs, you only need to write one thunk function to handle all three.

mlfFeval() calls the thunk function through the pointer it retrieves from the built-in table, passing it a pointer to the function to be executed, the number of input and output arguments, and the input and output argument arrays. Thunk functions also use the mlfFeval() calling convention.

The thunk function then translates from the calling convention used by mlfFeval() (arrays of arguments) to the standard C Math Library calling convention (an explicit list of arguments), executes the function, and returns the results to mlfFeval().

How Function-Functions Use mlfFeval( )

Extending the mlfFeval() Table