comp.lang.ruby

Robert Klemme

4/16/2006 10:21:00 AM

James Herdman wrote:
> I'm reading the last little fragment of the Pick Ax book during my exam
> study breaks right now, and I'm contemplating this UnboundMethod hotness.
>
> Pick Ax says, "As with aliases, unbounch methods are references to the
> definition of the method at the time they are created" with the example
>
> [CODE BEGIN]
> unbound_length = String.instance_method(:length)
> class String
> def length
> 99
> end
> end
> str = "cat"
> str.length -> 99
> bound_length = unbound_length.bind(str)
> bound_length.call -> 3
> [CODE END]
>
> The first part of my question is if I can read "... are references to
> the definition of the definition of the method at the time they are
> created" as being equivalent to "copy of"? (My assumption is yes, given
> that my unbound copy isn't altered by redefinition #length.)

IMHO not. In the example there is an initial version of the method
"length". A reference to that is kept alive with "unbound_method".
Then there is a second version of the method created by the class body
that replaces the original def (but that is still accessible via
"unbound_length"). If you execute instance_method twice in the
beginning you get two UnboundMethod objects but both point to the same
piece of code.

> The second part of my question what about the contents of
> unbound_length. Does it actually contain code, or is it just a memory
> location? (I.E. is there a way I can have unbound_length spit out the
> code for #length?)

No. There is no standard mechanism to convert some ruby object back
into the source code. You'll have to parse the code yourself.

Kind regards

robert