From maintainers-request at octave dot org Sun Oct 30 06:41:11 2005
Subject: Function pointers in subfunctions
From: Andy Adler <adler at site dot uOttawa dot ca>
To: octave-maintainers at bevo dot che dot wisc dot edu
Date: Sun, 30 Oct 2005 07:39:42 -0500 (EST)

There appears to be a limitation for octave's function pointers,
when used with the octave-forge optimization functions.

For example, this works:
    function testfcn()
       disp( fmins( at fcn, 0, [], [], 2, 1) )

       function y= fcn( x, p1, p2)
          y= abs( p1*x + p2 );

    octave:28> testfcn
    -0.50000


but this doesn't
    function testfcn()
       disp( fmins( at fcn, 0, [], [], 2, 1) )

       function y= fcn( x, p1, p2)
          y= ptr_fcn(x,p1,p2);

       function y= ptr_fcn( x, p1, p2)
          y= abs( p1*x + p2 );

    octave:27> testfcn
    error: `ptr_fcn' undefined near line 6 column 7
    error: evaluating assignment expression near line 6, column 5
    error: called from `testfcn:fcn' in file `.... /testfcn.m'

Both work in Matlab.

As an aside, I would really like to be able to use
'fsolve' with parameters provided. I tried to look at
fsolve.cc but it seemed to be based on a really old
method of creating tmp functions, so I wanted to check
whether it was worth trying to submit a patch to fsolve.cc
or not.

--
Andy Adler <adler at site dot uOttawa dot ca> 1(613)562-5800x6218