//////////////////////////////////////////////////////////////////////
                               rvs 0.6.1
	retroactive versioning system
	          a versioning system that allows you to check
	            in commit 2 before commit 1
//////////////////////////////////////////////////////////////////////

introduction

   The name is a little bit of a nod to RCS (revision control
   system), and even less to CVS. I'm not particularly fond of CVS,
   but recognize it's impact on the scm scene.

   rvs is about being able to go back and edit *anything* that has
   already been committed. Yes, some other SCMs do this, but fuck
   things proper if someone else has a copy of the old version.

   Why would you want to do this? Take for example the Bazaar
   repository rvs is hosted on: It starts at 0.6.0, what if I later
   want to import 0.0.0 - 0.5.10? Or if I'm trying to construct a
   history of BSD, and import all the BSDs I can find, such as
   1,3,4, then latter find 2?

system requirements

   Pretty much any *nix system should be able to run rvs. If you
   need to use a shell other than GNU bash, run ./configure with
   the --SHELL=YOUR_SHELL option. rvs is not designed for Windows,
   but should be able to run in bash if you have some *nix pack
   installed.

   To my knowledge, this is the absolute requirements (all need to
   be in your PATH):
     * GNU bash
     * cat      (included in GNU Core Utils)
     * cp       (included in GNU Core Utils)
     * cut      (included in GNU Core Utils)
     * echo     (included in GNU Core Utils)
     * mkdir    (included in GNU Core Utils)
     * rm       (included in GNU Core Utils)
     * sed      (included in GNU Core Utils)
     * sha1sum  (included in GNU Core Utils)
     * tempfile (included in GNU Core Utils)
     * tr       (included in GNU Core Utils)

building

   rvs doesn't exactly use the GNU build system, but acts much like
   it does. To build rvs with default configuration, simply run:
   	$ ./configure
   	$ make
   	# make install

   It's generally considered good practice to build in another
   directory than the source directory. This is not nescessary in
   rvs , all the created file are put in another directory anyway.
   If you would still like to do this, it can be done in rvs-0.6.1
   and up.

  Configuration

   There are several configuration variables that can be set. The
   default values are kept in the file `Variables'.

   Variables is in the format `VAR_NAME<tab>VALUE'. You can modify
   these configuration variables by editing the `Variables' file
   directly, or by running ./configure such as:
   	$ ./configure --VAR_NAME=VALUE

   The different configuration variables are as follows:
   VER   	value='0.6.0'
   	rvs's internal varsion number
   SHELL 	value='/usr/bin/env bash'
   	the shell that scripts will be run in.
   prefix	value="$HOME"
   	where the files will be installed. For me the defualt is
   	`/home/luke'. Many of you will want to change this to '/usr'
   bindir	value='bin'
   	the binaries folder inside the prefix. If I leave the prefix
   	and bindir the same, for me rvs is installed in
   	`/home/luke/bin'. If I change prefix to '/usr', then rvs will
   	be installed in `/usr/bin'
   libdir	value='/etc/rvs/'
   	where the rvs libraries will be installed. In rvs almost
   	nothing is actually done in the execurable you call when you
   	type `rvs' ([$prefix]/[$bindir]/rvs) but option handling. All
   	the real work is done in modular sub-programs. I call them
   	libraries, even though they are independend programs that
   	communicate via pipes.  Note that they probably should NOT be
   	located in your PATH.

~ Luke Shumaker <LukeShu@sbcglobal.net>
Happy Hacking!