Basics

Factory methods

Future and Event are created indirectly with constructor methods in FactoryMethods. They are not designed for inheritance but rather for composition.

Concurrent::Promises::FactoryMethods.instance_methods(false)
# => [:zip,
#     :create,
#     :delay,
#     :future,
#     :resolvable_future,
#     :resolvable_event,
#     :resolvable_event_on,
#     :resolvable_future_on,
#     :future_on,
#     :resolved_future,
#     :fulfilled_future,
#     :rejected_future,
#     :resolved_event,
#     :delay_on,
#     :schedule,
#     :schedule_on,
#     :zip_futures,
#     :zip_futures_on,
#     :zip_events,
#     :zip_events_on,
#     :any_resolved_future,
#     :any_resolved_future_on,
#     :any,
#     :any_fulfilled_future,
#     :any_fulfilled_future_on,
#     :any_event,
#     :any_event_on]

The module can be included or extended where needed.

Class.new do
  include Concurrent::Promises::FactoryMethods

  def a_method
    resolvable_event
  end
end.new.a_method
# => <#Concurrent::Promises::ResolvableEvent:0x7fb5e090d308 pending>

mod = Module.new do
  extend Concurrent::Promises::FactoryMethods
end 
mod.resolvable_event
# => <#Concurrent::Promises::ResolvableEvent:0x7fb5e0907750 pending>

The default executor can be changed by overriding default_executor method inherited from Concurrent::Promises::FactoryMethods.

mod = Module.new do
  extend Concurrent::Promises::FactoryMethods
  def self.default_executor
    :fast
  end
end 
mod.future { 1 }.default_executor        # => :fast
Concurrent::Promises.future { 1 }.default_executor
# => :io

The module is already extended into Concurrent::Promises for convenience.

Concurrent::Promises.resolvable_event
# => <#Concurrent::Promises::ResolvableEvent:0x7fb5e1a6c680 pending>

Asynchronous task

The most basic use-case of the framework is asynchronous processing. A task can be processed asynchronously by using a future factory method. The block will be executed on an internal thread pool.

Arguments of future are passed to the block and evaluation starts immediately.

future = Concurrent::Promises.future(0.1) do |duration|
  sleep duration
  :result
end
# => <#Concurrent::Promises::Future:0x7fb5e1a64a70 pending>

Asks if the future is resolved, here it will be still in the middle of the sleep call.

future.resolved?                         # => false

Retrieving the value will block until the future is resolved.

future.value                             # => :result
future.resolved?                         # => true

If the task fails we talk about the future being rejected.

future = Concurrent::Promises.future { raise 'Boom' }
# => <#Concurrent::Promises::Future:0x7fb5e22ca930 rejected>

There is no result, the future was rejected with a reason.

future.value                             # => nil
future.reason                            # => #<RuntimeError: Boom>

It can be forced to raise the reason for rejection when retrieving the value.

begin
  future.value! 
rescue => e 
  e
end                                      # => #<RuntimeError: Boom>

Which is the same as future.value! rescue $! which will be used hereafter.

Or it can be used directly as argument for raise, since it implements exception method.

raise future rescue $!                   # => #<RuntimeError: Boom>

States

Lets define a inspection helper for methods.

def inspect_methods(*methods, of:)
  methods.reduce({}) { |h, m| h.update m => of.send(m) }
end

Event has pending and resolved state. 

event = Concurrent::Promises.resolvable_event 
inspect_methods(:state, :pending?, :resolved?, of: event)
# => {:state=>:pending, :pending?=>true, :resolved?=>false}

event.resolve 
inspect_methods(:state, :pending?, :resolved?, of: event)
# => {:state=>:resolved, :pending?=>false, :resolved?=>true}

Future's resolved state is further specified to be fulfilled or rejected.

future = Concurrent::Promises.resolvable_future 
inspect_methods(:state, :pending?, :resolved?, :fulfilled?, :rejected?, 
    of: future)
# => {:state=>:pending,
#     :pending?=>true,
#     :resolved?=>false,
#     :fulfilled?=>false,
#     :rejected?=>false}

future.fulfill :value 
inspect_methods(:state, :pending?, :resolved?, :fulfilled?, :rejected?,
    :result, :value, :reason, of: future)
# => {:state=>:fulfilled,
#     :pending?=>false,
#     :resolved?=>true,
#     :fulfilled?=>true,
#     :rejected?=>false,
#     :result=>[true, :value, nil],
#     :value=>:value,
#     :reason=>nil}

future = Concurrent::Promises.rejected_future StandardError.new 
inspect_methods(:state, :pending?, :resolved?, :fulfilled?, :rejected?, 
    :result, :value, :reason, of: future)
# => {:state=>:rejected,
#     :pending?=>false,
#     :resolved?=>true,
#     :fulfilled?=>false,
#     :rejected?=>true,
#     :result=>[false, nil, #<StandardError: StandardError>],
#     :value=>nil,
#     :reason=>#<StandardError: StandardError>}

Direct creation of resolved futures

When an existing value has to wrapped in a future it does not have to go through evaluation as follows.

Concurrent::Promises.future { :value }
# => <#Concurrent::Promises::Future:0x7fb5e22a05b8 pending>

Instead it can be created directly.

Concurrent::Promises.fulfilled_future(:value)
# => <#Concurrent::Promises::Future:0x7fb5e38052c8 fulfilled>
Concurrent::Promises.rejected_future(StandardError.new('Ups'))
# => <#Concurrent::Promises::Future:0x7fb5e08f7be8 rejected>
Concurrent::Promises.resolved_future(true, :value, nil)
# => <#Concurrent::Promises::Future:0x7fb5e08f61a8 fulfilled>
Concurrent::Promises.resolved_future(false, nil, StandardError.new('Ups'))
# => <#Concurrent::Promises::Future:0x7fb5e08f4150 rejected>

Chaining

Big advantage of promises is ability to chain tasks together without blocking current thread.

Concurrent::Promises.
    future(2) { |v| v.succ }.
    then(&:succ).
    value!                               # => 4

As future factory method takes argument, then method takes as well. Any supplied arguments are passed to the block, and the library ensures that they are visible to the block.

Concurrent::Promises.
    future('3') { |s| s.to_i }.
    then(2) { |v, arg| v + arg }.
    value                                # => 5
Concurrent::Promises.
    fulfilled_future('3').
    then(&:to_i).
    then(2, &:+).
    value                                # => 5
Concurrent::Promises.
    fulfilled_future(1).
    chain(2) { |fulfilled, value, reason, arg| value + arg }.
    value                                # => 3

Passing the arguments in (similarly as for a thread Thread.new(arg) { |arg| do_stuff arg }) is required, both following examples may break.

arg = 1                                  # => 1
Thread.new { do_stuff arg }
# => #<Thread:0x0000000000000002@promises.in.md:203 run>
Concurrent::Promises.future { do_stuff arg }
# => <#Concurrent::Promises::Future:0x7fb5e11cd010 pending>

Branching, and zipping

Besides chaining it can also be branched.

head    = Concurrent::Promises.fulfilled_future -1 
branch1 = head.then(&:abs) 
branch2 = head.then(&:succ).then(&:succ) 

branch1.value!                           # => 1
branch2.value!                           # => 1

It can be combined back to one future by zipping (zip, &).

branch1.zip(branch2).value!              # => [1, 1]
(branch1 & branch2).
    then { |a, b| a + b }.
    value!                               # => 2
(branch1 & branch2).
    then(&:+).
    value!                               # => 2
Concurrent::Promises.
    zip(branch1, branch2, branch1).
    then { |*values| values.reduce(&:+) }.
    value!                               # => 3

Instead of zipping only the first one can be taken if needed.

Concurrent::Promises.any(branch1, branch2).value!
# => 1
(branch1 | branch2).value!               # => 1

Blocking methods

In these examples we have used blocking methods like value extensively for their convenience, however in practice is better to avoid them and continue chaining.

If they need to be used (e.g. when integrating with threads), value! is a better option over value when rejections are not dealt with differently. Otherwise the rejection are not handled and probably silently forgotten.

Error handling

When one of the tasks in the chain fails, the rejection propagates down the chain without executing the tasks created with then.

Concurrent::Promises.
    fulfilled_future(Object.new).
    then(&:succ).
    then(&:succ).
    result
# => [false,
#     nil,
#     #<NoMethodError: undefined method `succ' for #<Object:0x0000000000000003>>]

As then chained tasks execute only on fulfilled futures, there is a rescue method which chains a task which is executed only when the future is rejected. It can be used to recover from rejection.

Using rescue to fulfill to 0 instead of the error.

Concurrent::Promises.
    fulfilled_future(Object.new).
    then(&:succ).
    then(&:succ).
    rescue { |err| 0 }.
    result                               # => [true, 0, nil]

Rescue not executed when there is no rejection.

Concurrent::Promises.
    fulfilled_future(1).
    then(&:succ).
    then(&:succ).
    rescue { |e| 0 }. 
    result                               # => [true, 3, nil]

Tasks added with chain are evaluated always.

Concurrent::Promises.
    fulfilled_future(1).
    chain { |fulfilled, value, reason| fulfilled ? value : reason }.
    value!                               # => 1
Concurrent::Promises.
    rejected_future(StandardError.new('Ups')).
    chain { |fulfilled, value, reason| fulfilled ? value : reason }.
    value!                               # => #<StandardError: Ups>

Zip is rejected if any of the zipped futures is.

rejected_zip = Concurrent::Promises.zip(
    Concurrent::Promises.fulfilled_future(1),
    Concurrent::Promises.rejected_future(StandardError.new('Ups')))
# => <#Concurrent::Promises::Future:0x7fb5e19c6640 rejected>
rejected_zip.result
# => [false, [1, nil], [nil, #<StandardError: Ups>]]
rejected_zip.
    rescue { |reason1, reason2| (reason1 || reason2).message }.
    value                                # => "Ups"

Delayed futures

Delayed futures will not evaluate until asked by touch or other method requiring resolution. 

future = Concurrent::Promises.delay { sleep 0.1; 'lazy' }
# => <#Concurrent::Promises::Future:0x7fb5e19b60d8 pending>
sleep 0.1 
future.resolved?                         # => false
future.touch
# => <#Concurrent::Promises::Future:0x7fb5e19b60d8 pending>
sleep 0.2 
future.resolved?                         # => true

All blocking methods like wait, value call touch and trigger evaluation.

Concurrent::Promises.delay { :value }.value
# => :value

It propagates trough chain up allowing whole or partial lazy chains.

head    = Concurrent::Promises.delay { 1 } 
branch1 = head.then(&:succ) 
branch2 = head.delay.then(&:succ) 
join    = branch1 & branch2 

sleep 0.1

Nothing resolves.

[head, branch1, branch2, join].map(&:resolved?)
# => [false, false, false, false]

Force branch1 evaluation.

branch1.value                            # => 2
sleep 0.1 
[head, branch1, branch2, join].map(&:resolved?)
# => [true, true, false, false]

Force evaluation of both by calling value on join.

join.value                               # => [2, 2]
[head, branch1, branch2, join].map(&:resolved?)
# => [true, true, true, true]

Flatting

Sometimes it is needed to wait for a inner future. Apparent solution is to wait inside the future Concurrent::Promises.future { Concurrent::Promises.future { 1+1 }.value }.value however as mentioned before, value calls should be avoided to avoid blocking threads. Therefore there is a flat method which is a correct solution in this situation and does not block any thread.

Concurrent::Promises.future { Concurrent::Promises.future { 1+1 } }.flat.value!
# => 2

A more complicated example.

Concurrent::Promises.
    future { Concurrent::Promises.future { Concurrent::Promises.future { 1 + 1 } } }.
    flat(1).
    then { |future| future.then(&:succ) }.
    flat(1).
    value!                               # => 3

Scheduling

Tasks can be planned to be executed with a time delay.

Schedule task to be executed in 0.1 seconds.

scheduled = Concurrent::Promises.schedule(0.1) { 1 }
# => <#Concurrent::Promises::Future:0x7fb5e118f670 pending>
scheduled.resolved?                      # => false

Value will become available after 0.1 seconds. 

scheduled.value                          # => 1

It can be used in the chain as well, where the delay is counted form a moment its parent resolves. Therefore following future will be resolved in 0.2 seconds.

future = Concurrent::Promises.
    future { sleep 0.1; :result }.
    schedule(0.1).
    then(&:to_s).
    value!                               # => "result"

Time can be used as well.

Concurrent::Promises.schedule(Time.now + 10) { :val }
# => <#Concurrent::Promises::Future:0x7fb5e1868cd0 pending>

Resolvable Future and Event:

Sometimes it is required to resolve a future externally, in these cases resolvable_future and resolvable_event factory methods can be uses. See Concurrent::Promises::ResolvableFuture and Concurrent::Promises::ResolvableEvent.

future = Concurrent::Promises.resolvable_future
# => <#Concurrent::Promises::ResolvableFuture:0x7fb5e225bf80 pending>

The thread will be blocked until the future is resolved

thread = Thread.new { future.value } 
future.fulfill 1
# => <#Concurrent::Promises::ResolvableFuture:0x7fb5e225bf80 fulfilled>
thread.value                             # => 1

Future can be resolved only once.

future.fulfill 1 rescue $!
# => #<Concurrent::MultipleAssignmentError: Future can be resolved only once. It's [true, 1, nil], trying to set [true, 1, nil]. {:current_result=>[true, 1, nil], :new_result=>[true, 1, nil]}>
future.fulfill 2, false                  # => false

How are promises executed?

Promises use global pools to execute the tasks. Therefore each task may run on different thread which implies that users have to be careful not to depend on Thread local variables (or they have to set at the begging of the task and cleaned up at the end of the task).

Since the tasks are running on may different threads of the thread pool, it's better to follow following rules:

TODO: This part to be extended

Advanced

Callbacks

queue  = Queue.new                       # => #<Thread::Queue:0x0000000000000004>
future = Concurrent::Promises.delay { 1 + 1 }
# => <#Concurrent::Promises::Future:0x7fb5e222a610 pending>

future.on_fulfillment { queue << 1 } # evaluated asynchronously
future.on_fulfillment! { queue << 2 } # evaluated on resolving thread

queue.empty?                             # => true
future.value                             # => 2
queue.pop                                # => 2
queue.pop                                # => 1

Using executors

Factory methods, chain, and callback methods have all other version of them which takes executor argument.

It takes an instance of an executor or a symbol which is a shortcuts for the two global pools in concurrent-ruby. fast for short and non-blocking tasks and :io for blocking and long tasks.

Concurrent::Promises.future_on(:fast) { 2 }.
    then_on(:io) { File.read __FILE__ }.
    value.size                           # => 26689

Run (simulated process)

Similar to flatting is running. When run is called on a future it will flat indefinitely as long the future fulfils into a Future value. It can be used to simulate a thread like processing without actually occupying the thread.

count = lambda do |v|
  v += 1
  v < 5 ? Concurrent::Promises.future_on(:fast, v, &count) : v
end
# => #<Proc:0x0000000000000005@promises.in.md:511 (lambda)>
400.times.
    map { Concurrent::Promises.future_on(:fast, 0, &count).run.value! }.
    all? { |v| v == 5 }                  # => true

Therefore the above example finished fine on the the :fast thread pool even though it has much less threads than there is the simulated process.

Interoperability

Actors

Create an actor which takes received numbers and returns the number squared. 

actor = Concurrent::Actor::Utils::AdHoc.spawn :square do
  -> v { v ** 2 }
end
# => #<Concurrent::Actor::Reference:0x7fb5e08190f0 /square (Concurrent::Actor::Utils::AdHoc)>

Send result of 1+1 to the actor, and add 2 to the result send back from the actor.

Concurrent::Promises.
    future { 1 + 1 }.
    then_ask(actor).
    then { |v| v + 2 }.
    value!                               # => 6

So (1 + 1)**2 + 2 = 6.

The ask method returns future.

actor.ask(2).then(&:succ).value!         # => 5

Channel

There is an implementation of channel as well. Lets start by creating a channel with capacity 2 messages.

ch1 = Concurrent::Promises::Channel.new 2
# => <#Concurrent::Promises::Channel:0x7fb5e1a74128 size:2>

We push 3 messages, it can be observed that the last future representing the push is not fulfilled since the capacity prevents it. When the work which fills the channel depends on the futures created by push it can be used to create back pressure – the filling work is delayed until the channel has space for more messages.

pushes = 3.times.map { |i| ch1.push i }
# => [<#Concurrent::Promises::Future:0x7fb5e1a6d3a0 fulfilled>,
#     <#Concurrent::Promises::Future:0x7fb5e1a6cae0 fulfilled>,
#     <#Concurrent::Promises::Future:0x7fb5e1a67ab8 pending>]
ch1.pop.value!                           # => 0
pushes
# => [<#Concurrent::Promises::Future:0x7fb5e1a6d3a0 fulfilled>,
#     <#Concurrent::Promises::Future:0x7fb5e1a6cae0 fulfilled>,
#     <#Concurrent::Promises::Future:0x7fb5e1a67ab8 fulfilled>]

A selection over channels can be created with select_channel factory method. It will be fulfilled with a first message available in any of the channels. It returns a pair to be able to find out which channel had the message available.

ch2    = Concurrent::Promises::Channel.new 2
# => <#Concurrent::Promises::Channel:0x7fb5e22c8bd0 size:2>
result = Concurrent::Promises.select_channel(ch1, ch2)
# => <#Concurrent::Promises::ResolvableFuture:0x7fb5e22c3d38 fulfilled>
result.value!
# => [<#Concurrent::Promises::Channel:0x7fb5e1a74128 size:2>, 1]

Concurrent::Promises.future { 1+1 }.then_push_channel(ch1)
# => <#Concurrent::Promises::Future:0x7fb5e1a5d860 pending>
result = (
    Concurrent::Promises.fulfilled_future('%02d') &      
        Concurrent::Promises.select_channel(ch1, ch2)).
    then { |format, (channel, value)| format format, value }
# => <#Concurrent::Promises::Future:0x7fb5e22b9f90 pending>
result.value!                            # => "02"

ProcessingActor

There is also a new implementation of actors based on the Channel and the ability of promises to simulate process. The actor runs as a process but also does not occupy a thread per actor as previous Concurrent::Actor implementation. This implementation is close to Erlang actors, therefore OTP can be ported for this actors (and it's planned).

The simplest actor is a one which just computes without even receiving a message.

actor = Concurrent::ProcessingActor.act(an_argument = 2) do |actor, number|
  number ** 3
end
# => <#Concurrent::ProcessingActor:0x7fb5e1a54490 termination:pending>
actor.termination.value!                 # => 8

Let's receive some messages though.

add_2_messages = Concurrent::ProcessingActor.act do |actor|
  # Receive two messages then terminate normally with the sum.
  (actor.receive & actor.receive).then do |a, b|
    a + b
  end
end
add_2_messages.tell 1
# => <#Concurrent::Promises::Future:0x7fb5e1a44518 pending>
add_2_messages.termination.resolved?     # => false
add_2_messages.tell 3
# => <#Concurrent::Promises::Future:0x7fb5e22a9e10 pending>
add_2_messages.termination.value!        # => 4

Actors can also be used to apply back pressure to a producer. Let's start by defining an actor which a mailbox of size 2.

slow_counter = -> (actor, count) do
  actor.receive.then do |command, number|
    sleep 0.1
    case command
    when :add
      slow_counter.call actor, count + number
    when :done
      # terminate
      count
    end
  end
end

actor = Concurrent::ProcessingActor.act_listening( 
    Concurrent::Promises::Channel.new(2), 
    0,
    &slow_counter)
# => <#Concurrent::ProcessingActor:0x7fb5e11dd960 termination:pending>

Now we can create a producer which will push messages only when there is a space available in the mailbox. We use promises to free a thread during waiting on a free space in the mailbox.

produce = -> receiver, i do
  if i < 10
    receiver.
        # send a message to the actor, resolves only after the message is 
        # accepted by the actor's mailbox
        tell([:add, i]).
        # send incremented message when the above message is accepted 
        then(i+1, &produce)
  else
    receiver.tell(:done)
    # do not continue 
  end
end

Concurrent::Promises.future(actor, 0, &produce).run.wait!
# => <#Concurrent::Promises::Future:0x7fb5e1a1fd80 fulfilled>

actor.termination.value!                 # => 45

Use-cases

Simple background processing

Concurrent::Promises.future { do_stuff }
# => <#Concurrent::Promises::Future:0x7fb5e19cff88 pending>

Parallel background processing

tasks = 4.times.map { |i| Concurrent::Promises.future(i) { |i| i*2 } }
# => [<#Concurrent::Promises::Future:0x7fb5e19cc450 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e229b428 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e229a0a0 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e2298480 pending>]
Concurrent::Promises.zip(*tasks).value!
# => [0, 2, 4, 6]

Actor background processing

Actors are mainly keep and isolate state, they should stay responsive not being blocked by a longer running computations. It desirable to offload the work to stateless promises.

Lets define an actor which will process jobs, while staying responsive, and tracking the number of tasks being processed.

class Computer < Concurrent::Actor::RestartingContext
  def initialize
    super()
    @jobs = {}
  end

  def on_message(msg)
    command, *args = msg
    case command
    # new job to process
    when :run
      job        = args[0]
      @jobs[job] = envelope.future
      # Process asynchronously and send message back when done.
      Concurrent::Promises.future(&job).chain(job) do |fulfilled, value, reason, job|
        self.tell [:done, job, fulfilled, value, reason]
      end
      # Do not make return value of this method to be answer of this message.
      # We are answering later in :done by resolving the future kept in @jobs.
      Concurrent::Actor::Behaviour::MESSAGE_PROCESSED
    when :done
      job, fulfilled, value, reason = *args
      future                        = @jobs.delete job
      # Answer the job's result.
      future.resolve fulfilled, value, reason
    when :status
      { running_jobs: @jobs.size }
    else
      # Continue to fail with unknown message.
      pass 
    end
  end
end

Create the computer actor and send it 3 jobs.

computer = Concurrent::Actor.spawn Computer, :computer
# => #<Concurrent::Actor::Reference:0x7fb5e11a6b90 /computer (Computer)>
results = 3.times.map { computer.ask [:run, -> { sleep 0.1; :result }] }
# => [<#Concurrent::Promises::Future:0x7fb5e1873d38 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e1872cd0 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e1871a88 pending>]
computer.ask(:status).value!             # => {:running_jobs=>3}
results.map(&:value!)                    # => [:result, :result, :result]

Solving the Thread count limit by thread simulation

Sometimes an application requires to process a lot of tasks concurrently. If the number of concurrent tasks is high enough than it is not possible to create a Thread for each of them. A partially satisfactory solution could be to use Fibers, but that solution locks the application on MRI since other Ruby implementations are using threads for each Fiber.

This library provides a Concurrent::Promises::Future#run method on a future to simulate threads without actually accepting one all the time. The run method is similar to Concurrent::Promises::Future#flat but it will keep flattening until it's fulfilled with non future value, then the value is taken as a result of the process simulated by run.

body = lambda do |v|
  # Some computation step of the process    
  new_v = v + 1
  # Is the process finished?
  if new_v < 5
    # Continue computing with new value, does not have to be recursive.
    # It just has to return a future.
    Concurrent::Promises.future(new_v, &body)
  else
    # The process is finished, fulfill the final value with `new_v`.
    new_v
  end
end
Concurrent::Promises.future(0, &body).run.value! # => 5

This solution works well an any Ruby implementation.

TODO: More examples to be added.

Cancellation

Simple

Lets have two processes which will count until cancelled.

source, token = Concurrent::Cancellation.create
# => [<#Concurrent::Cancellation:0x7fb5e182ea80 canceled:false>,
#     <#Concurrent::Cancellation::Token:0x7fb5e182dbf8 canceled:false>]

count_until_cancelled = -> token, count do
  if token.canceled?
    count
  else
    Concurrent::Promises.future token, count+1, &count_until_cancelled
  end
end 

futures = Array.new(2) do
  Concurrent::Promises.future(token, 0, &count_until_cancelled).run
end
# => [<#Concurrent::Promises::Future:0x7fb5e1826f88 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e18249b8 pending>]

sleep 0.01 
source.cancel                            # => true
futures.map(&:value!)                    # => [53, 53]

Cancellation can also be used as event or future to log or plan re-execution.

token.to_event.chain do
  # log cancellation
  # plane re-execution
end

Parallel background processing with cancellation

Each task tries to count to 1000 but there is a randomly failing test. The tasks share a cancellation, when one of them fails it cancels the others.

source, token = Concurrent::Cancellation.create
# => [<#Concurrent::Cancellation:0x7fb5e22abdf0 canceled:false>,
#     <#Concurrent::Cancellation::Token:0x7fb5e22ab5f8 canceled:false>]
tasks = 4.times.map do |i|
  Concurrent::Promises.future(source, token, i) do |source, token, i|
    count = 0
    1000.times do
      break count = :cancelled if token.canceled?
      count += 1
      sleep 0.01
      if rand > 0.95
        source.cancel
        raise 'random error'
      end
      count
    end
  end
end
# => [<#Concurrent::Promises::Future:0x7fb5e22a81c8 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e22a3218 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e22a1d28 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e22a0658 pending>]
Concurrent::Promises.zip(*tasks).result
# => [false,
#     [nil, :cancelled, nil, :cancelled],
#     [#<RuntimeError: random error>, nil, #<RuntimeError: random error>, nil]]

Without the randomly failing part it produces following.

source, token = Concurrent::Cancellation.create
# => [<#Concurrent::Cancellation:0x7fb5e1a2cd00 canceled:false>,
#     <#Concurrent::Cancellation::Token:0x7fb5e1a2c2d8 canceled:false>]
tasks = 4.times.map do |i|
  Concurrent::Promises.future(source, token, i) do |source, token, i|
    count = 0
    1000.times do
      break count = :cancelled if token.canceled?
      count += 1
      # sleep 0.01
      # if rand > 0.95
      #   source.cancel
      #   raise 'random error'
      # end
    end
    count
  end
end
Concurrent::Promises.zip(*tasks).result
# => [true, [1000, 1000, 1000, 1000], nil]

Throttling concurrency

By creating an actor managing the resource we can control how many threads is accessing the resource. In this case one at the time.

data      = Array.new(10) { |i| '*' * i }
# => ["",
#     "*",
#     "**",
#     "***",
#     "****",
#     "*****",
#     "******",
#     "*******",
#     "********",
#     "*********"]
DB = Concurrent::Actor::Utils::AdHoc.spawn :db, data do |data|
  lambda do |message|
    # pretending that this queries a DB
    data[message]
  end
end

concurrent_jobs = 11.times.map do |v|
  DB.
      # ask the DB with the `v`, only one at the time, rest is parallel
      ask(v).
      # get size of the string, rejects for 11
      then(&:size).
      # translate error to a value (message of the exception)
      rescue { |reason| reason.message } 
end 

Concurrent::Promises.zip(*concurrent_jobs).value!
# => [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, "undefined method `size' for nil:NilClass"]

Often there is more then one DB connections, then the pool can be used.

pool_size = 5                            # => 5

DB_POOL = Concurrent::Actor::Utils::Pool.spawn!('DB-pool', pool_size) do |index|
  # DB connection constructor
  Concurrent::Actor::Utils::AdHoc.spawn(
      name: "connection-#{index}", 
      args: [data]) do |data|
    lambda do |message|
      # pretending that this queries a DB
      data[message]
    end
  end
end

concurrent_jobs = 11.times.map do |v|
  DB_POOL.
      # ask the DB with the `v`, only one at the time, rest is parallel
      ask(v).
      # get size of the string, rejects for 11
      then(&:size).
      # translate error to a value (message of the exception)
      rescue { |reason| reason.message } 
end 

Concurrent::Promises.zip(*concurrent_jobs).value!
# => [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, "undefined method `size' for nil:NilClass"]

In other cases the DB adapter maintains its internal connection pool and we just need to limit concurrent access to the DB's API to avoid the calls being blocked.

Lets pretend that the #[] method on DB_INTERNAL_POOL is using the internal pool of size 3. We create throttle with the same size

DB_INTERNAL_POOL = Concurrent::Array.new data 
# => ["",
#     "*",
#     "**",
#     "***",
#     "****",
#     "*****",
#     "******",
#     "*******",
#     "********",
#     "*********"]

max_tree = Concurrent::Throttle.new 3
# => <#Concurrent::Throttle:0x7fb5e21d3360 limit:3 can_run:3>

futures = 11.times.map do |i|
  max_tree.
      # throttled tasks, at most 3 simultaneous calls of [] on the database
      throttled_future { DB_INTERNAL_POOL[i] }.
      # un-throttled tasks, unlimited concurrency
      then { |starts| starts.size }.
      rescue { |reason| reason.message }
end 

futures.map(&:value!)
# => [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, "undefined method `size' for nil:NilClass"]

Long stream of tasks, applying back pressure

Lets assume that we queuing an API for a data and the queries can be faster than we are able to process them. This example shows how to use channel as a buffer and how to apply back pressure to slow down the queries. 

require 'json' 

channel       = Concurrent::Promises::Channel.new 6
# => <#Concurrent::Promises::Channel:0x7fb5e11e4e90 size:6>
source, token = Concurrent::Cancellation.create
# => [<#Concurrent::Cancellation:0x7fb5e11de9a0 canceled:false>,
#     <#Concurrent::Cancellation::Token:0x7fb5e11ddc58 canceled:false>]

def query_random_text(token, channel)
  Concurrent::Promises.future do
    # for simplicity the query is omitted
    # url = 'some api'
    # Net::HTTP.get(URI(url))
    sleep 0.1
    { 'message' => 
        'Lorem ipsum rhoncus scelerisque vulputate diam inceptos' 
    }.to_json
  end.then(token) do |value, token|
    # The push to channel is fulfilled only after the message is successfully
    # published to the channel, therefore it will not continue querying until 
    # current message is pushed.
    channel.push(value) | 
        # It could wait on the push indefinitely if the token is not checked 
        # here with `or` (the pipe).
        token.to_future
  end.flat_future.then(token) do |_, token|
    # query again after the message is pushed to buffer
    query_random_text(token, channel) unless token.canceled?
  end
end

words          = []                      # => []
words_throttle = Concurrent::Throttle.new 1
# => <#Concurrent::Throttle:0x7fb5e1a37d90 limit:1 can_run:1>

def count_words_in_random_text(token, channel, words, words_throttle)
  channel.pop.then do |response|
    string = JSON.load(response)['message']
    # processing is slower than querying
    sleep 0.2
    words_count = string.scan(/\w+/).size
  end.then_throttled_by(words_throttle, words) do |words_count, words|
    # safe since throttled to only 1 task at a time
    words << words_count
  end.then(token) do |_, token|
    # count words in next message
    unless token.canceled?
      count_words_in_random_text(token, channel, words, words_throttle)
    end
  end
end

query_processes = 3.times.map do
  Concurrent::Promises.future(token, channel, &method(:query_random_text)).run
end
# => [<#Concurrent::Promises::Future:0x7fb5e1a2e128 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e1a279b8 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e1a25820 pending>]

word_counter_processes = 2.times.map do
  Concurrent::Promises.future(token, channel, words, words_throttle, 
      &method(:count_words_in_random_text)).run
end
# => [<#Concurrent::Promises::Future:0x7fb5e11ccf70 pending>,
#     <#Concurrent::Promises::Future:0x7fb5e11c6b48 pending>]

sleep 0.5                                # => 1

Let it run for a while then cancel it and ensure that the runs all fulfilled (therefore ended) after the cancellation. Finally print the result.

source.cancel                            # => true
query_processes.map(&:wait!) 
# => [<#Concurrent::Promises::Future:0x7fb5e1a2e128 fulfilled>,
#     <#Concurrent::Promises::Future:0x7fb5e1a279b8 fulfilled>,
#     <#Concurrent::Promises::Future:0x7fb5e1a25820 fulfilled>]
word_counter_processes.map(&:wait!)
# => [<#Concurrent::Promises::Future:0x7fb5e11ccf70 fulfilled>,
#     <#Concurrent::Promises::Future:0x7fb5e11c6b48 fulfilled>]
words                                    # => [7, 7, 7, 7]

Compared to using threads directly this is highly configurable and compostable solution.

Periodic task

By combining schedule, run and Cancellation periodically executed task can be easily created.

repeating_scheduled_task = -> interval, token, task do
  Concurrent::Promises.
      # Schedule the task.
      schedule(interval, token, &task).
      # If successful schedule again. 
      # Alternatively use chain to schedule always.
      then { repeating_scheduled_task.call(interval, token, task) }
end

cancellation, token = Concurrent::Cancellation.create
# => [<#Concurrent::Cancellation:0x7fb5e222a7c8 canceled:false>,
#     <#Concurrent::Cancellation::Token:0x7fb5e2228b80 canceled:false>]

task = -> token do
  5.times do
    token.raise_if_canceled
    # do stuff
    print '.'
    sleep 0.01
  end
end

result = Concurrent::Promises.future(0.1, token, task, &repeating_scheduled_task).run
# => <#Concurrent::Promises::Future:0x7fb5e1987968 pending>
sleep 0.2                                # => 0
cancellation.cancel                      # => true
result.result
# => [false,
#     nil,
#     #<Concurrent::CancelledOperationError: Concurrent::CancelledOperationError>]