From: "Gray, Jeff" <jeff.gray@...>
Date: 2002-05-30T08:43:00+09:00
Subject: Parsing of equals sign as part of method name

We all know that one can either define a method that has a name of the form
'method=' (or use attr_accessor/attr_writer), and when Ruby sees 

  obj.method = value

 the equivalent of

  obj.method=(value)

I'm finding that the parser never expects or allows more than one argument
to the 'method=' style of method naming.  For example, you could define a
method interface of:

  def method=(arg, &block)

but the parser doesn't want to let you pass in a block to this call.  Both
of the following generate parse errors:

  obj.method = 12 { <some_code_here> }
  obj.method=(12) { <some_code_here> }

(Simply removing the '=' from the method name allows the second case to
work.)  I couldn't find anything in the pickaxe book or in ruby-talk
archives about this restriction.  I'm on Ruby version 1.6.5 -- has this
behavior changed in later versions, or is it a permanent trait of Ruby
syntax?

Thanks,

  - jeff