Is it possible to pass no arguments in order to invoke default values for functions compiled into a C++ shared library with the MATLAB Compiler?

Question

MathWorks Support Team 2010-2-25

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/96067-is-it-possible-to-pass-no-arguments-in-order-to-invoke-default-values-for-functions-compiled-into-a

I would like to create a C++ shared library using the MATLAB Compiler that I can call from my C++ program. I would like to be able to use default values in my shared library function when no argument is supplied. In MATLAB, I can use the NARGIN function to detect when arguments are not specified.

For example:

function c = increment_matrix(a,b)
% Returns c = a + b.
% If b is not specified, then c = a + 1.
% If a and b are not specified, then c = 0.
if nargin < 2, b = 1; end
if nargin < 1, a = -1; end
c = a + b;

When I compile this into a C++ shared library with the MATLAB Compiler, it creates a function signature requiring two input arguments, but I would still like to be able to use the function with one or no input arguments.

请先登录，再回答此问题。

Answer 1

MathWorks Support Team 2010-2-25

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/96067-is-it-possible-to-pass-no-arguments-in-order-to-invoke-default-values-for-functions-compiled-into-a#answer_105419

Because the MATLAB Compiler creates a function signature that matches the one in the MATLAB file, you cannot call a library function with fewer or more arguments than is defined in the function.

You can work around this as follows:

1. Add ISEMPTY conditions to your MATLAB file, so that you can also pass empty variables into your function. These empty variables will be replaced by your default values.

For example:

function c = increment_matrix(a,b)
% Returns c = a + b.
% If b is not specified, then c = a + 1.
% If a and b are not specified, then c = 0.
% if nargin < 2, b = 1; end               % OLD
% if nargin < 1, a = -1; end              % OLD
if nargin < 2 || isempty(b), b = 1; end   % NEW
if nargin < 1 || isempty(a), a = -1; end  % NEW
c = a + b;

2. In your C++ program, create an empty mwArray to use for passing to your functions.

For example:

mwArray empty_input_argument(0, 0, mxDOUBLE_CLASS, mxREAL);

3. In your C++ program use the empty input argument (from Step 2) to emulate a function call with fewer arguments.

For example:

 increment_matrix(1,out1,in1,in2);  // Both input arguments specified
 increment_matrix(1,out1,in1,empty_input_argument);  // One input argument specified.
 increment_matrix(1,out1,empty_input_argument,empty_input_argument);  // No input arguments specified.

The following files are attached to help illustrate this.

- increment_matrix.m

The source code for the increment_matrix function used in this solution.

- libincrmtrx.dll

The C++ shared library that was generated with the MATLAB Compiler by issuing the following command:
mcc -W cpplib:libincrmtrx -T link:lib increment_matrix.m -v

- libincrmtrx.ctf

The CTF archive file that was also generated from the same command issued to create libincrmtrx.dll above.

- call_increment_matrix.cpp

The source code to a sample C++ program that calls the increment_matrix function from the libincrmtrx C++ shared library.

- call_increment_matrix.exe

The compiled sample C++ program, which requires libincrmtrx.dll and libincrmtrx.ctf to be in the same directory during execution.  The program should produce the following output:
 The value of 'increment_matrix(1,2)' is:
 3
 The value of 'increment_matrix(1)' is:
 2
 The value of 'increment_matrix()' is:
 0