Ractor.new

Ractor.new(*args, name: nil) {|*args| block } → ractor

# File tmp/rubies/ruby-3.3.0/ractor.rb, line 273
def self.new(*args, name: nil, &block)
  b = block # TODO: builtin bug
  raise ArgumentError, "must be called with a block" unless block
  if __builtin_cexpr!("RBOOL(ruby_single_main_ractor)")
    warn("Ractor is experimental, and the behavior may change in future versions of Ruby! " \
         "Also there are many implementation issues.", uplevel: 0, category: :experimental)
  end
  loc = caller_locations(1, 1).first
  loc = "#{loc.path}:#{loc.lineno}"
  __builtin_ractor_create(loc, name, args, b)
end

Create a new Ractor with args and a block.

The given block (Proc) will be isolated (can’t access any outer variables). self inside the block will refer to the current Ractor.

Example

r = Ractor.new { puts "Hi, I am #{self.inspect}" }
r.take
# Prints "Hi, I am #<Ractor:#2 test.rb:1 running>"

Any args passed are propagated to the block arguments by the same rules as objects sent via send/Ractor.receive. If an argument in args is not shareable, it will be copied (via deep cloning, which might be inefficient).

Example

arg = [1, 2, 3]
puts "Passing: #{arg} (##{arg.object_id})"
r = Ractor.new(arg) {|received_arg|
  puts "Received: #{received_arg} (##{received_arg.object_id})"
}
r.take
# Prints:
#   Passing: [1, 2, 3] (#280)
#   Received: [1, 2, 3] (#300)

Ractor’s name can be set for debugging purposes:

Example

r = Ractor.new(name: 'my ractor') {}; r.take
p r
#=> #<Ractor:#3 my ractor test.rb:1 terminated>

See also .make_shareable .receive .main .receive_if also in 4.1 · 4.0 · 3.4 · 3.2 · 3.1 · 3.0

‹ prev.make_shareable next ›.receive