How to pass blocks between methods in Ruby < 3.0.0 without using &block arguments.

gistfile1.rb

# UPDATE: The ability to call Proc.new without a block was removed in Ruby 3.0.0
# See https://bugs.ruby-lang.org/issues/10499 for more information.

# So you might have heard that this is slow:
#
#   def some_method(&block)
#     block.call
#   end
#
# Compared to:
#
#   def some_method
#     yield
#   end
#
# But how can you pass a block from one method to another without
# using the &block argument?
#
#   def some_method(&block)
#     some_other_method(&block)
#   end
#
# Behold!

def some_method
  some_other_method(&Proc.new)
end

def some_other_method
  puts yield * 2
end

some_method { "Magic!" }
# => "Magic!Magic!"

# As mentioned by Aaron Patterson in his RubyConf X presentation
# "ZOMG WHY IS THIS CODE SO SLOW", this means that you can avoid the cost
# of instantiating a Proc object altogether if suitable.

def some_method
  if block_given?
    some_other_method(&Proc.new)
  end
end

# The key to this is in Proc.new which will create a Proc object from the
# block passed to the current method when called without any arguments of
# its own.

# For more information, see the following:
# http://www.ruby-doc.org/core/classes/Proc.html#M000547
# http://confreaks.net/videos/427-rubyconf2010-zomg-why-is-this-code-so-slow (at around 30:00)

Thanks, this is exactly what I needed to DRY out some code. :)

Thanks Bro this save me a lot of troubles

Interesting!
What if you have block as an optional argument? With &Proc.new it will raise, so here using &block is a must right?

For example:
some_other_method(arg1, arg2, &Proc.new)
And if you call some_method without a block it still should execute.

@zauzaj I wrote about this in a bit more detail in "Passing Blocks in Ruby Without &block" but, yes, you'd need to detect whether an optional block was passed with block_given? before calling Proc.new in that situation.

The code does not work with Ruby-3.0.1. The error is "tried to create Proc object without a block (ArgumentError)"

@kopylovvlad you're right, it looks like this was changed in Ruby 3.0.0 (see #10499 and #15554) but the documentation still describes the old behaviour though this also looks like it has been fixed.

I've added a comment to the gist to flag this change of behaviour.