avram/README

avram - applicative virtual machine code interpreter

Copyright (C) 2006,2007,2012 Dennis Furey

This package includes a virtual machine code interpreter and a
development library. It enables you to run programs that have been
compiled into the virtual code format that it uses. A compiler to
generate code in this format is available separately at
http://gueststar.github.com/ursala. 

The quick install guide (assuming you have root):
-----------------------

Change to the directory with this README file in it, and do the
following.

configure
make
su
make install

If this doesn't work, try reading INSTALL, and please let me know what
you had to do to get it to work so I can include it in this file or
improve the configuration script.

Installation for non-root users (whose admins are busy):
-------------------------------

Change to the directory with this README file in it and type

configure --prefix=$HOME
make
make install

The avram binary will be installed under $HOME/bin, which you need to
put in your path. You might also want to modify your $MANPATH and your
$INFOPATH environment variables to include $HOME/man and $HOME/info.
You may substitute any directory you own for $HOME.

See below if you want to link avram with external math libraries,
which will need to be installed first.

Installation for non-root users with small disk quotas (aka students):
------------------------------------------------------

Unpack the tarball in /tmp, which is world writable on most systems,
cd to /tmp/avram-0.3.0 (or current version) and type

configure --prefix=/tmp
make
strip src/avram

The third line is optional but recommended. When it's finished
building, copy src/avram to your home directory wherever you keep your
executables (probably ~/bin). You won't have any on-line documentation
with this kind of installation, but you can always read it on the home
page. Nor will you be able to link your own code against the avram
library, but you'll use only a few hundred kbytes for the executable
and will be able to execute unmodified virtual code scripts if they're
done right.

A combination of the above installations is to follow the building in
/tmp with make install, and then copy /tmp/lib to your home directory
as well as /tmp/bin/avram. That will give you the executable and the
library but no documentation.

Installation of Expect
----------------------

Avram has cooler features when built on a machine with the expect
library, available at http://expect.sourceforge.net. If your system
doesn't have any files named /usr/lib/*expect* or
/usr/local/lib/*expect*, you might want to install the expect library
in order to get the full benefit before installing avram. Version
numbers shown below may vary.

1) download tcl8.5.12-src.tar.gz and expect5.45.tar.gz
2) tar -zxvf tcl8.5.12-src.tar.gz
3) tar -zxvf expect5.45.tar.gz
4) cd tcl8.5.12/unix
5) su
6) configure; make; make install
7) cd ../../expect-5.45
8) configure; make; make install

It is possible to install these libraries non-root by using configure
--prefix=$HOME. As a show of my solidarity with non-root users, I've
made the avram configure script look under $HOME for shared libraries
and header files. You will then have to put your $HOME/lib directory
in your LD_LIBRARY_PATH environment variable.

Installation of other libraries
-------------------------------

Some other libraries with which avram is able to interface may be of
interest. These should be installed prior to compiling avram, if you
want, but avram will also work without them.

* gsl      integration, differentiation, and series acceleration
* mpfr     arbitrary precision arithmetic
* minpack  non-linear optimization
* lapack   linear algebra
* fftw     fast fourier transforms
* Rmath    statistical and transcendental functions
* ufsparse sparse matrices
* glpk     linear programming by the simplex method

Keep an eye on the output messages from the configuration script to
ensure the libraries you want have been detected if you're compiling
avram yourself, and then as a check execute "avram -e" with the newly
built binary.

Comments are welcome.

	- Dennis Furey <[email protected]>