Re: C language interface to octave (eng* function library)

[A. Scottedward Hodel]

Are you aware of the .oct file support for dynamically linked
functions?  There are several examples of these in the octave
distribution in
    examples/oregonator.cc
    examples/hello.cc
and in the 
    src/DLD-FUNCTIONS 
directory (with some straightforward editing of the include file paths 
for mkoctfile).  I've used this resource to develop an update of the qz 
function in Octave to solve generalized eigenvalues (to appear in a later 
release; requires a few extra fortran routines, etc.)  I've also
implemented some examples of .oct files working with structures
in the octave controls toolbox (see ftp://ftp.eng.auburn.edu/pub/hodel), 
which is also undergoing some rewriting at this time (maintaining
compatibility with previous releases!).

I found it somewhat less painful to write .oct files than I did
mex-files (although I had more experience working with this sort
of interface, so it's not a fair comparison).  The .oct files have
access to the full Octave library, and so you'll already have access
to much of BLAS, LAPACK, etc. when writing your C++ code (which can
easily act as a wrapper to call your own C-code if you wish).

Those specific functions you list in your note are proprietary to
MATLAB; I don't think anyone has written direct interfaces to them.
Backward compatibility and portability among versions of matlab (4,5) 
may be a problem to hinder such an implementation, since Octave ranges 
over a wide range of target machines.

A S Hodel Dept Elect Eng, Auburn Univ,AL  36849-5201
On leave at NASA Marshall Space Flight Center (205) 544-1426

----------
>From: Jesse Bennett <address@hidden>
>To: address@hidden
>Subject: C language interface to octave (eng* function library)
>Date: Thu, Jul 23, 1998, 4:22 PM
[edited]
>Is there any octave support (or ongoing development) for an Octave
>interface library?  I find that Matlab is very effective as a user
>interface for much of the R&D that I am involved in, and I am interested
>in the possibility of using octave for this work.  In other words, is
>there any compatibility with the libmat.a functions eng*?  E.g.,