From: Rick DeNatale <rick.denatale@...>
Date: 2009-04-23T22:44:25+09:00
Subject: Re: kernal::binding problem

On Thu, Apr 23, 2009 at 3:09 AM, Joel VanderWerf
<vjoel@path.berkeley.edu> wrote:
> Tom Cloyd wrote:

>>
>> eval( "moduleName.methodName", binding)
>>
>> Causes *crash* due to by attempt to access a variable which exists in the
>> calling environment but not in the module. In spite of passing the binding,
>> the method knows nothing of the variable. It's not working.
>>
>
> You've got something like this, right?
>
> module M
>  def M.meth
>    p x
>  end
> end
>
> x = 1
> eval "M.meth", binding # Undefined variable x
>
>
> The binding that you pass into eval here only affects variables that are
> unbound in the string. In this case there are none. The x in M.meth is a
> bound variable in the scope of the method.

And to stress the obvious

eval "some string"

which uses the current binding, is exactly equivalent to

eval "some string", binding

Assuming that binding is a call to Kernel#binding and not a local variable.

So

eval "M.meth"
and
eval "M.meth", binding

will both do the same thing, including throwing any exceptions.

-- 
Rick DeNatale

Blog: http://talklikeaduck.denhaven2.com/
Twitter: http://twitter.com/RickDeNatale
WWR: http://www.workingwithrails.com/person/9021-rick-denatale
LinkedIn: http://www.linkedin.com/in/rickdenatale