Using the C++ Math Library

Handling Exceptions

The MATLAB C++ Math Library delivers error messages via exceptions. This section:

• Provides a brief overview of C++ exception-handling
• Describes how to handle C++ Math Library exceptions in your application
• Describes how you can customize exception handling by replace the default exception handling routines with routines of your own design.
• Describes how the MATLAB C++ Math Library implements exceptions, and how you can use this mechanism to throw and catch exceptions in your own code.

Refer to Appendix C for a list of MATLAB C++ error messages.

C++ Exception Handling Overview

Many earlier error-handling schemes reported errors via a return value from a function. That mechanism was inconvenient and unreliable for two reasons. First, it did not allow function composition, where one function call is nested in the argument list of another. For example, f(g(x)) composes f() and g(). Second, the scheme placed the burden for checking error codes on the programmer.

Many other schemes, including the one used by the standard C library, use a global variable in place of returned error codes. This mechanism solves one problem, function composition, but still requires that the programmer check for errors.

The C++ exception-handling mechanism suffers from neither of these problems. Exception-handling does not require that each function return an error code, which means that functions can be composed. In addition, exceptions cannot be ignored by a programmer because an uncaught exception terminates the program. If a programmer forgets to handle an exception, abrupt program termination is a potent reminder.

Output to a GUI

Handling C++ Math Library Exceptions in Your Code