From: Robert Klemme <shortcutter@...>
Date: 2006-08-21T23:40:12+09:00
Subject: Re: working around "single-pass" constraint

On 21.08.2006 16:29, Rich Morin wrote:
> I like to order my program code files as follows:
> 
>   Initial comment header
> 
>   require statements, etc.
> 
>   "main" code
> 
>   method definitions
> 
> Unfortunately, Ruby has a "single pass" execution flow
> (a bit reminiscent of Pascal's single-pass compiler :-/)
> that doesn't allow this order to be used, by default.
> 
> 
> My current hack is to wrap the "main" code in a method
> (e.g., main :-) and invoke it at the end of the file, I
> can get around this issue.  I also like the fact that it
> (a) puts the code at the same level of indentation as the
> code in my methods and (b) limits the scope of variables:
> 
>   #!/usr/bin/env ruby -w
>   #
>   # foo - code formatting demo
> 
>   def main
>     bar
>   end
> 
>   def bar
>     puts 'Hi!'
>   end
> 
>   main
> 
> 
> However, I have two concerns with this approach:
> 
>   *  I might be missing a common Ruby idiom (at least,
>      common within the small group of folks who like to
>      order their code as I do).

As I don't belong to that group I don't have any insight there. :-)

Personally I don't mind having the order Ruby imposes.

>   *  I might be in danger of running into some sort of
>      "dynamic language issue", in which some method has
>      to be defined early, because it gets used in main's
>      _definition_.
> 
>      I don't know, unfortunately, whether this is a real
>      concern.  Or, if it is, whether I could work around
>      it by putting the location-sensitive definitions in
>      front of main's definition.

This is not an issue at all:

16:34:42 [~]: ruby -e 'def main() foo() end; def foo() puts "icks" end; 
main()'
icks

IOW, the method needs to be there when it's called - not earlier.

Kind regards

	robert