Different data types in/out of functions.

[Sydney]

Hi, I want to write a function(to adjust settings increment or decrement them), I would want to carry the current setting into the function, then return the adjusted value. I have different variables, int8's, signed and unsigned int16's so the question is do I need a different function for each variable, or how do I do it with one? Thanks.

[Ttelmah]

First thing is that for many 'related' types, a single function may well work fine. For example, if you had a function to decrement a number, and wrote it to handle an int32, then called it with an int8 value, the latter will automatically be converted to an int32, the function will execute, and return an int32, which if you put the result into an int8, will automatically be converted back to an int8.
The second part, is that if the function needs to behave differently with different sized values, then you can use function overloading. With this, you write two (or more) different functions, with exactly the same name, but with each declared for a different variable type. When a function is called, the compiler automatically checks the variable type, and calls the function that matches that type.
The third is to manually override the operation of a single function, by declaring it to accept a pointer to your data, and a value saying what the type of the data is. So if you declared an enum, with values for int8, int16, int32 etc., and then called a function, with a pointer to an intxx value, and this enum, then you can test in the target function, what type the data being pointed to is, by looking at the enum, and adjust the code accordingly.
Which will suit your target best, will depend on how the code needs to change with different data types.

Best Wishes

[Wayne_]

Function overloading wasn't introduced until C++, does CCS C support it ?

Ttelmah is correct in that you could declare a function to take a pointer of type void and cast it, the problem is that the function won't know what type the original var is without passing another parameter to specify the type as Ttelmah stated.

Another option is to use a variable argument function definition as this can take multiple arguments of different types. http://publications.gbdirect.co.uk/c_book/chapter9/stdarg.html explains it.

[Ttelmah]

Yes. Function overloading is supported in CCS.
Function overloading is in ANSI C, it appeared befoe C++.
V4, gives quite a few parts of ANSI compatibility (with various degrees of success), but this actually worked qute early.
It is used quite a bit in the V4 code, for things like the maths library, where there are 64bit versions for the PCD compiler, that replace the 32bit versions, if 64bit variables are passed.

Best Wishes

[Sydney]

Thanks guys, seems to be working fine using the 1st method