This more accurately reflects what it controls

Issues were two fold:
- HistoryItem statement.raw was getting mangled for multiline commands due to macro-related changes in _input_line_to_statement()
- HistoryItem pretty printing wasn't using rstrip() anymore in verbose mode

I added a couple unit tests in the process of getting here.  But we should add some explicit unit tests of _input_line_to_statement() for cases like:
- basic single-line command
- macro single-line command
- multiline command

allow_cli_args is now an argument to __init__ instead of a cmd2 class member

cmd2.Cmd.cmdloop() now returns self.exit_code which should be an integer

Also:
- Refactored examples to call sys.exit(app.cmdloop()) in their __main__
- Running transcript tests now sets the exit_code accordingly based on success/failure
- Updated CHANGELOG
- Updated README
- Updated Sphinx docs
- Added unit test for case when transcript test fails

The following are now arguments to cmd2.Cmd.__init__() instead of class attributes:
* allow_redirection
* multiline_commands
* terminators
* shortcuts

Added a couple read-only properties for convenience of cmd2.Cmd accessing immutable members from self.statement_parser