From: ahoward <ahoward@...>
Date: 2003-01-31T01:02:35+09:00
Subject: Re: Local variables & blocks

On Thu, 30 Jan 2003, Yukihiro Matsumoto wrote:

> Describe your feeling and ideas.

please also see prior threads.

  summary :

    there is only ONE common case of people wanting to export variable from a
    block, and that is that one of some type of accumulator living on beyond
    the scope of a block, the

      a = []
      loop do
	...
	a << method
	...
      end

    situation.

    i'm not saying that's not a good reason to do something, or that there are
    not other situation where it might make good sense to export a variable,
    but it is the exception, _not_ the rule

    that why i suggested the _exceptional_ case of

      loop do
	export a = []
	...
	a << method
	...
      end

    export -> declare/define, once only, a variable in the parent's scope
	      (dynamic)

    as solving that problem.  though someone pointed out it was more typing
    than the current method of 'a = []' before the loop i really _do_ think
    this idiom could lead to more elegant _looking_ code whose intent is
    clearer.

> This is why I post.  What kind of situations do you think this change does
> not cover properly?

  * Threads
  * GC
  * Comfort in knowing a module writer cannot fill your namespace up
    inadvertantly
  * Need to track memory usage in blocks :

      (2 << 15).times do |i|
	eval "leak_#{i} = 'a big fat leak' * (2 << 16)"
      end

but perhaps there is a was around these issues i've overlooked (very likey -
i' not that smart!) ;-)

-a


-- 

 ====================================
 | Ara Howard
 | NOAA Forecast Systems Laboratory
 | Information and Technology Services
 | Data Systems Group
 | R/FST 325 Broadway
 | Boulder, CO 80305-3328
 | Email: ahoward@fsl.noaa.gov
 | Phone:  303-497-7238
 | Fax:    303-497-7259
 ====================================