Re: Getting a function to return an array



"Andrew Gentile" <andrewkgentile@xxxxxxxxx> wrote in message
news:1167694481.265614.212540@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Hello,
I am trying to write a function which takes several floats as
inputs, then returns 1x6 array of floats. I normally program in
Matlab, and I am often confused by C. It seems that this should be an
easy task, but I can't seem to get it done. Can someone please
include a brief snippet of code which demonstrates how to get a
function to accept 2 floats as inputs, and returns an array of 6
floats.

This is an interface design question, and there are a couple of ways of
doing it.

I hesitate to give this advice, as 'C' gives you more than enough rope to
hang yourself.

METHOD 1--ALLOCATE THE MEMORY IN THE CALLER

void caller(void)
{
float inputs[2];
float outputs[6];

.... assign the inputs, inputs[0] and inputs [1].

callee(inputs, outputs);

.... Here the outputs are assigned by the called function.
}

void callee(float *inputs, float *outputs)
{
Here you may use inputs[0] and inputs[1] (but no more).

Here you may assign outputs[0] through outputs[5] (but no more).
}

On second thought, I won't describe alternate methods.

In 'C', the brackets [] are an operator. Although the compiler may make
some optimizations, the brackets just say "index into this array, pointed
to, by a number of elements".

There really is no notion in 'C' of going beyond the limits of the array.
That has to be controlled carefully. Allocation information is not packaged
with the array.

In the example, above, addressing outputs[6] (beyond the end of the array)
is very likely to be a run-time disaster.

There are various "safer" ways to handle the above, but the above comes down
to the fact that the caller and callee have to have agreement about what is
being passed. They both need to agree that inputs is only 2 floats big and
outputs is only 6 floats big. The "unsafe" way to have that agreement is
implicit ... with the information not explicitly passed. The above code is
unsafe, but it should work for you as long as you don't go beyond the array
boundaries.

Also, inputs and outputs are passed with the "&" operator because 'C' has
the convention that arrays and pointers are the same thing. For other data
types, a & would be required.

Please post back with any specific questions.

Dave.


.



Relevant Pages

  • Re: converting to bits
    ... > Does C have some handy functions to convert chars, ints and floats to bit ... what type of array would you want to store ...
    (comp.lang.c)
  • Re: List<t> generic type - Fastest way to copy into an unmanaged array?
    ... > floats into an unmanaged array of floats. ... > As far as the managed collection is implemented, ... > boxing/unboxing issues. ...
    (microsoft.public.dotnet.languages.vc)
  • List<t> generic type - Fastest way to copy into an unmanaged array?
    ... floats into an unmanaged array of floats. ... As far as the managed collection is implemented, ... fast way of copying a tradional managed array into an unmanaged array. ...
    (microsoft.public.dotnet.languages.vc)
  • Re: optimizing simple-arrays
    ... I'd expect your get-floatN routines to ... have to perform the array lookup at each reference, ... You might do better, also, to have maxvalue as single floats ... > function was the exact bottleneck, so I just optimized the whole thing ...
    (comp.lang.lisp)