NAG CPP Interface
Change Log

• •Minor change to a callback interface in handle_​solve_​ipopt.
  Old specification:

  void hess(const std::vector<double> &x, const nagcpp::types::f77_integer idf,
   const double sigma, const std::vector<double> &lamda,
   const nagcpp::types::f77_integer nnzh, std::vector<double> &hx,
   nagcpp::types::f77_integer inform)
   

  New specification:

  void hess(const std::vector<double> &x, const nagcpp::types::f77_integer idf,
   const double sigma, const std::vector<double> &lamda,
   std::vector<double> &hx,
   nagcpp::types::f77_integer inform)
   

• •The number of Lagrange multipliers in handle_​solve_​ipopt is now inferred from the corresponding array.
  Old call:

  std::vector<double> u;
  ...
  handle_solve_ipopt(handle, ..., x, nnzu, u, ...);
   

  New call:

  std::vector<double> u(nnzu);
  ...
  handle_solve_ipopt(handle, ..., x, u, ...);
   

• • Enumerations under ArgIntent were renamed from IN, OUT and INOUT to IntentIN, IntentOUT and IntentINOUT respectively. This to avoid issues encountered when using Visual Studio in conjunction with another package that wants to define IN and OUT.

• •Readme converted to markdown.
• •Recommended Windows compilation switch for exceptions modified to /EHs.
• •Include files cosmetically regenerated for Mark 27.2.

• •Minor changes to the example results for ex_e04kf, ex_e04mt, ex_e04rb and ex_e04st.
• •Include files regenerated for Mark 27.1.
• •New option LPIPM Max Iterative Refinement for handle_​solve_​lp_​ipm.
• •Minor changes to the callback interfaces for handle_​solve_​bounds_​foas and handle_​solve_​ipopt:
  handle_​solve_​bounds_​foas

  Sample old call:

  void objfun(const std::vector<double> &x, double &fx,
   const nagcpp::types::f77_integer inform)
  

  New call:

  void objfun(const std::vector<double> &x, double &fx,
   nagcpp::types::f77_integer inform)
  

  Old call:

  void objgrd(const std::vector<double> &x, std::vector<double> &fdx,
   const nagcpp::types::f77_integer inform)
  

  New call:

  void objgrd(const std::vector<double> &x, std::vector<double> &fdx,
   nagcpp::types::f77_integer inform)
  

  handle_​solve_​ipopt

  Sample old call:

  void confun(const std::vector<double> &x,
   const nagcpp::types::f77_integer ncnln, std::vector<double> &gx)
  

  New call:

  void confun(const std::vector<double> &x,
   const nagcpp::types::f77_integer ncnln, std::vector<double> &gx,
   nagcpp::types::f77_integer inform)
  

  Old call:

  void congrd(const std::vector<double> &x,
   const nagcpp::types::f77_integer nnzgd, std::vector<double> &gdx)
  

  New call:

  void congrd(const std::vector<double> &x,
   std::vector<double> &gdx, const nagcpp::types::f77_integer nnzgd)
  

  Old call:

  void hess(const std::vector<double> &x, const nagcpp::types::f77_integer idf,
   const double sigma, const std::vector<double> &lamda,
   const nagcpp::types::f77_integer nnzh, std::vector<double> &hx)
  

  New call:

  void hess(const std::vector<double> &x, const nagcpp::types::f77_integer idf,
   const double sigma, const std::vector<double> &lamda,
   const nagcpp::types::f77_integer nnzh, std::vector<double> &hx,
   nagcpp::types::f77_integer inform)
  

Fixes

• •Enumeration data_handling::INOUT was renamed to ArgIntent to avoid issues encountered when using Visual Studio.
• •Various minor changes to a number of header files to give a clean build log when using Visual Studio.
• •Various minor changes to code comments in a number of header files.
• •Various minor changes to coding style.

Enhancements

• ‘makefile’ changed to enable use of cl.exe when using WSL.