I'm reading the Pickaxe and on page 391, there is a code listing for
the class method `once` from the Date module in Ruby:
def once(*ids) #:nodoc:
for id in ids
module_eval <<-"end:"
alias_method :__#{id.to_i}__, :#{id.to_s}
private :__#{id.to_i}__
def #{id.to_s}(*args, &block)
puts "Inside once: \#{args.inspect}"
(@__#{id.to_i}__ ||= [__#{id.to_i}__(*args, &block)])[0]
end
end:
end
end
I was wondering why the original method called is wrapped in an Array.
So I started digging the the source history hoping to find an
explanation. I found the commit that changed it to use the Array but
no explanation for the change. The commit where it was modified is
here:
I created a simple class and use the code with and without the Array
and it seems to work the same for my simple test. Just curious if I'm
not seeing something obvious.
They wrap it in an array so you that they can cache "nil" and "false"
return values:
result ||= expensive_computation
will still run the expensive computation if it returned nil or false,
however, if you did:
(result ||= [expensive_computation])[0],
you could cache nil as a return value and it wouldn't actually run the
expensive computation.
Matt Haley wrote:
···
I'm reading the Pickaxe and on page 391, there is a code listing for
the class method `once` from the Date module in Ruby:
def once(*ids) #:nodoc:
for id in ids
module_eval <<-"end:"
alias_method :__#{id.to_i}__, :#{id.to_s}
private :__#{id.to_i}__
def #{id.to_s}(*args, &block)
puts "Inside once: \#{args.inspect}"
(@__#{id.to_i}__ ||= [__#{id.to_i}__(*args, &block)])[0]
end
end:
end
end
I was wondering why the original method called is wrapped in an Array.
So I started digging the the source history hoping to find an
explanation. I found the commit that changed it to use the Array but
no explanation for the change. The commit where it was modified is
here:
I created a simple class and use the code with and without the Array
and it seems to work the same for my simple test. Just curious if I'm
not seeing something obvious.