module Concurrent
{include:file:README.md}
@!macro thread_safe_variable_comparison
## Thread-safe Variable Classes Each of the thread-safe variable classes is designed to solve a different problem. In general: * *{Concurrent::Agent}:* Shared, mutable variable providing independent, uncoordinated, *asynchronous* change of individual values. Best used when the value will undergo frequent, complex updates. Suitable when the result of an update does not need to be known immediately. * *{Concurrent::Atom}:* Shared, mutable variable providing independent, uncoordinated, *synchronous* change of individual values. Best used when the value will undergo frequent reads but only occasional, though complex, updates. Suitable when the result of an update must be known immediately. * *{Concurrent::AtomicReference}:* A simple object reference that can be updated atomically. Updates are synchronous but fast. Best used when updates a simple set operations. Not suitable when updates are complex. {Concurrent::AtomicBoolean} and {Concurrent::AtomicFixnum} are similar but optimized for the given data type. * *{Concurrent::Exchanger}:* Shared, stateless synchronization point. Used when two or more threads need to exchange data. The threads will pair then block on each other until the exchange is complete. * *{Concurrent::MVar}:* Shared synchronization point. Used when one thread must give a value to another, which must take the value. The threads will block on each other until the exchange is complete. * *{Concurrent::ThreadLocalVar}:* Shared, mutable, isolated variable which holds a different value for each thread which has access. Often used as an instance variable in objects which must maintain different state for different threads. * *{Concurrent::TVar}:* Shared, mutable variables which provide *coordinated*, *synchronous*, change of *many* stated. Used when multiple value must change together, in an all-or-nothing transaction.
TODO (pitr-ch 14-Mar-2017): deprecate, Future
, Promise
, etc.
Constants
- ArrayImplementation
@!macro internal_implementation_note
- AtomicBooleanImplementation
@!visibility private @!macro internal_implementation_note
- AtomicFixnumImplementation
@!visibility private @!macro internal_implementation_note
- AtomicReferenceImplementation
@!macro internal_implementation_note
- CancelledOperationError
Raised when an asynchronous operation is cancelled before execution.
- ConfigurationError
Raised when errors occur during configuration.
- CountDownLatchImplementation
@!visibility private @!macro internal_implementation_note
- Error
- ExchangerImplementation
@!visibility private @!macro internal_implementation_note
- GLOBAL_FAST_EXECUTOR
@!visibility private
- GLOBAL_IMMEDIATE_EXECUTOR
@!visibility private
- GLOBAL_IO_EXECUTOR
@!visibility private
- GLOBAL_LOGGER
@!visibility private
- GLOBAL_TIMER_SET
@!visibility private
- HashImplementation
@!macro internal_implementation_note
- IllegalOperationError
Raised when an operation is attempted which is not legal given the receiver’s current state
- ImmutabilityError
Raised when an attempt is made to violate an immutability guarantee.
- InitializationError
Raised when an object’s methods are called when it has not been properly initialized.
- LifecycleError
Raised when a lifecycle method (such as ‘stop`) is called in an improper sequence or when the object is in an inappropriate state.
- LockLocalVar
A ‘FiberLocalVar` is a variable where the value is different for each fiber. Each variable may have a default value, but when you modify the variable only the current fiber will ever see that change.
This is similar to Ruby’s built-in fiber-local variables (‘Thread.current`), but with these major advantages:
-
‘FiberLocalVar` has its own identity, it doesn’t need a Symbol.
-
Each Ruby’s built-in fiber-local variable leaks some memory forever (it’s a Symbol held forever on the fiber), so it’s only OK to create a small amount of them. ‘FiberLocalVar` has no such issue and it is fine to create many of them.
-
Ruby’s built-in fiber-local variables leak forever the value set on each fiber (unless set to nil explicitly). ‘FiberLocalVar` automatically removes the mapping for each fiber once the `FiberLocalVar` instance is GC’d.
@example
v = FiberLocalVar.new(14) v.value #=> 14 v.value = 2 v.value #=> 2
@example
v = FiberLocalVar.new(14) Fiber.new do v.value #=> 14 v.value = 1 v.value #=> 1 end.resume Fiber.new do v.value #=> 14 v.value = 2 v.value #=> 2 end.resume v.value #=> 14
-
- MaxRestartFrequencyError
Raised when an object with a start/stop lifecycle has been started an excessive number of times. Often used in conjunction with a restart policy or strategy.
- NULL
Various classes within allows for
nil
values to be stored, so a specialNULL
token is required to indicate the “nil-ness”. @!visibility private- NULL_LOGGER
Suppresses all output when used for logging.
- PromiseExecutionError
- RejectedExecutionError
Raised by an ‘Executor` when it is unable to process a given task, possibly because of a reject policy or other internal error.
- ResourceLimitError
Raised when any finite resource, such as a lock counter, exceeds its maximum limit/threshold.
- SemaphoreImplementation
@!visibility private @!macro internal_implementation_note
- SetImplementation
@!macro internal_implementation_note
- SingleThreadExecutorImplementation
- ThreadPoolExecutorImplementation
- TimeoutError
Raised when an operation times out.
- VERSION
Public Class Methods
Abort a currently running transaction - see ‘Concurrent::atomically`.
# File lib/concurrent-ruby/concurrent/tvar.rb, line 139 def abort_transaction raise Transaction::AbortError.new end
Run a block that reads and writes ‘TVar`s as a single atomic transaction. With respect to the value of `TVar` objects, the transaction is atomic, in that it either happens or it does not, consistent, in that the `TVar` objects involved will never enter an illegal state, and isolated, in that transactions never interfere with each other. You may recognise these properties from database transactions.
There are some very important and unusual semantics that you must be aware of:
-
Most importantly, the block that you pass to atomically may be executed
more than once. In most cases your code should be free of side-effects, except for via TVar.
-
If an exception escapes an atomically block it will abort the transaction.
-
It is undefined behaviour to use callcc or Fiber with atomically.
-
If you create a new thread within an atomically, it will not be part of
the transaction. Creating a thread counts as a side-effect.
Transactions within transactions are flattened to a single transaction.
@example
a = new TVar(100_000) b = new TVar(100) Concurrent::atomically do a.value -= 10 b.value += 10 end
# File lib/concurrent-ruby/concurrent/tvar.rb, line 82 def atomically raise ArgumentError.new('no block given') unless block_given? # Get the current transaction transaction = Transaction::current # Are we not already in a transaction (not nested)? if transaction.nil? # New transaction begin # Retry loop loop do # Create a new transaction transaction = Transaction.new Transaction::current = transaction # Run the block, aborting on exceptions begin result = yield rescue Transaction::AbortError => e transaction.abort result = Transaction::ABORTED rescue Transaction::LeaveError => e transaction.abort break result rescue => e transaction.abort raise e end # If we can commit, break out of the loop if result != Transaction::ABORTED if transaction.commit break result end end end ensure # Clear the current transaction Transaction::current = nil end else # Nested transaction - flatten it and just run the block yield end end
# File lib/concurrent-ruby/concurrent/dataflow.rb, line 56 def call_dataflow(method, executor, *inputs, &block) raise ArgumentError.new('an executor must be provided') if executor.nil? raise ArgumentError.new('no block given') unless block_given? unless inputs.all? { |input| input.is_a? IVar } raise ArgumentError.new("Not all dependencies are IVars.\nDependencies: #{ inputs.inspect }") end result = Future.new(executor: executor) do values = inputs.map { |input| input.send(method) } block.call(*values) end if inputs.empty? result.execute else counter = DependencyCounter.new(inputs.size) { result.execute } inputs.each do |input| input.add_observer counter end end result end
@return [Logger] Logger with provided level and output.
# File lib/concurrent-ruby/concurrent/concern/logging.rb, line 37 def self.create_simple_logger(level = Logger::FATAL, output = $stderr) # TODO (pitr-ch 24-Dec-2016): figure out why it had to be replaced, stdlogger was deadlocking lambda do |severity, progname, message = nil, &block| return false if severity < level message = block ? block.call : message formatted_message = case message when String message when Exception format "%s (%s)\n%s", message.message, message.class, (message.backtrace || []).join("\n") else message.inspect end output.print format "[%s] %5s -- %s: %s\n", Time.now.strftime('%Y-%m-%d %H:%M:%S.%L'), Logger::SEV_LABEL[severity], progname, formatted_message true end end
@return [Logger] Logger with provided level and output. @deprecated
# File lib/concurrent-ruby/concurrent/concern/logging.rb, line 69 def self.create_stdlib_logger(level = Logger::FATAL, output = $stderr) logger = Logger.new(output) logger.level = level logger.formatter = lambda do |severity, datetime, progname, msg| formatted_message = case msg when String msg when Exception format "%s (%s)\n%s", msg.message, msg.class, (msg.backtrace || []).join("\n") else msg.inspect end format "[%s] %5s -- %s: %s\n", datetime.strftime('%Y-%m-%d %H:%M:%S.%L'), severity, progname, formatted_message end lambda do |loglevel, progname, message = nil, &block| logger.add loglevel, message, progname, &block end end
Dataflow allows you to create a task that will be scheduled when all of its data dependencies are available. {include:file:docs-source/dataflow.md}
@param [Future] inputs zero or more ‘Future` operations that this dataflow depends upon
@yield The operation to perform once all the dependencies are met @yieldparam [Future] inputs each of the ‘Future` inputs to the dataflow @yieldreturn [Object] the result of the block operation
@return [Object] the result of all the operations
@raise [ArgumentError] if no block is given @raise [ArgumentError] if any of the inputs are not ‘IVar`s
# File lib/concurrent-ruby/concurrent/dataflow.rb, line 34 def dataflow(*inputs, &block) dataflow_with(Concurrent.global_io_executor, *inputs, &block) end
# File lib/concurrent-ruby/concurrent/dataflow.rb, line 44 def dataflow!(*inputs, &block) dataflow_with!(Concurrent.global_io_executor, *inputs, &block) end
# File lib/concurrent-ruby/concurrent/dataflow.rb, line 39 def dataflow_with(executor, *inputs, &block) call_dataflow(:value, executor, *inputs, &block) end
# File lib/concurrent-ruby/concurrent/dataflow.rb, line 49 def dataflow_with!(executor, *inputs, &block) call_dataflow(:value!, executor, *inputs, &block) end
Disables AtExit handlers including pool auto-termination handlers. When disabled it will be the application programmer’s responsibility to ensure that the handlers are shutdown properly prior to application exit by calling ‘AtExit.run` method.
@note this option should be needed only because of ‘at_exit` ordering
issues which may arise when running some of the testing frameworks. E.g. Minitest's test-suite runs itself in `at_exit` callback which executes after the pools are already terminated. Then auto termination needs to be disabled and called manually after test-suite ends.
@note This method should never be called
from within a gem. It should *only* be used from within the main application and even then it should be used only when necessary.
@deprecated Has no effect since it is no longer needed, see github.com/ruby-concurrency/concurrent-ruby/pull/841.
# File lib/concurrent-ruby/concurrent/configuration.rb, line 48 def self.disable_at_exit_handlers! deprecated "Method #disable_at_exit_handlers! has no effect since it is no longer needed, see https://github.com/ruby-concurrency/concurrent-ruby/pull/841." end
General access point to global executors. @param [Symbol, Executor] executor_identifier symbols:
- :fast - {Concurrent.global_fast_executor} - :io - {Concurrent.global_io_executor} - :immediate - {Concurrent.global_immediate_executor}
@return [Executor]
# File lib/concurrent-ruby/concurrent/configuration.rb, line 83 def self.executor(executor_identifier) Options.executor(executor_identifier) end
Global thread pool optimized for short, fast operations.
@return [ThreadPoolExecutor] the thread pool
# File lib/concurrent-ruby/concurrent/configuration.rb, line 55 def self.global_fast_executor GLOBAL_FAST_EXECUTOR.value! end
# File lib/concurrent-ruby/concurrent/configuration.rb, line 66 def self.global_immediate_executor GLOBAL_IMMEDIATE_EXECUTOR end
Global thread pool optimized for long, blocking (IO) tasks.
@return [ThreadPoolExecutor] the thread pool
# File lib/concurrent-ruby/concurrent/configuration.rb, line 62 def self.global_io_executor GLOBAL_IO_EXECUTOR.value! end
# File lib/concurrent-ruby/concurrent/concern/logging.rb, line 109 def self.global_logger GLOBAL_LOGGER.value end
# File lib/concurrent-ruby/concurrent/concern/logging.rb, line 113 def self.global_logger=(value) GLOBAL_LOGGER.value = value end
Global thread pool user for global timers.
@return [Concurrent::TimerSet] the thread pool
# File lib/concurrent-ruby/concurrent/configuration.rb, line 73 def self.global_timer_set GLOBAL_TIMER_SET.value! end
Leave a transaction without committing or aborting - see ‘Concurrent::atomically`.
# File lib/concurrent-ruby/concurrent/tvar.rb, line 144 def leave_transaction raise Transaction::LeaveError.new end
@!macro monotonic_get_time
Returns the current time as tracked by the application monotonic clock. @param [Symbol] unit the time unit to be returned, can be either :float_second, :float_millisecond, :float_microsecond, :second, :millisecond, :microsecond, or :nanosecond default to :float_second. @return [Float] The current monotonic time since some unspecified starting point @!macro monotonic_clock_warning
# File lib/concurrent-ruby/concurrent/utility/monotonic_time.rb, line 15 def monotonic_time(unit = :float_second) Process.clock_gettime(Process::CLOCK_MONOTONIC, unit) end
@!visibility private
# File lib/concurrent-ruby/concurrent/atomic/lock_local_var.rb, line 7 def self.mutex_owned_per_thread? return false if Concurrent.on_jruby? || Concurrent.on_truffleruby? mutex = Mutex.new # Lock the mutex: mutex.synchronize do # Check if the mutex is still owned in a child fiber: Fiber.new { mutex.owned? }.resume end end
# File lib/concurrent-ruby/concurrent/configuration.rb, line 87 def self.new_fast_executor(opts = {}) FixedThreadPool.new( [2, Concurrent.processor_count].max, auto_terminate: opts.fetch(:auto_terminate, true), idletime: 60, # 1 minute max_queue: 0, # unlimited fallback_policy: :abort, # shouldn't matter -- 0 max queue name: "fast" ) end
# File lib/concurrent-ruby/concurrent/configuration.rb, line 98 def self.new_io_executor(opts = {}) CachedThreadPool.new( auto_terminate: opts.fetch(:auto_terminate, true), fallback_policy: :abort, # shouldn't matter -- 0 max queue name: "io" ) end
Number of physical processor cores on the current system. For performance reasons the calculated value will be memoized on the first call.
On Windows the Win32 API will be queried for the ‘NumberOfCores from Win32_Processor`. This will return the total number “of cores for the current instance of the processor.” On Unix-like operating systems either the `hwprefs` or `sysctl` utility will be called in a subshell and the returned value will be used. In the rare case where none of these methods work or an exception is raised the function will simply return 1.
@return [Integer] number physical processor cores on the current system
@see github.com/grosser/parallel/blob/4fc8b89d08c7091fe0419ca8fba1ec3ce5a8d185/lib/parallel.rb
@see msdn.microsoft.com/en-us/library/aa394373(v=vs.85).aspx @see www.unix.com/man-page/osx/1/HWPREFS/ @see linux.die.net/man/8/sysctl
# File lib/concurrent-ruby/concurrent/utility/processor_counter.rb, line 107 def self.physical_processor_count processor_counter.physical_processor_count end
Number of processors seen by the OS and used for process scheduling. For performance reasons the calculated value will be memoized on the first call.
When running under JRuby the Java runtime call ‘java.lang.Runtime.getRuntime.availableProcessors` will be used. According to the Java documentation this “value may change during a particular invocation of the virtual machine… [applications] should therefore occasionally poll this property.” Subsequently the result will NOT be memoized under JRuby.
Otherwise Ruby’s Etc.nprocessors will be used.
@return [Integer] number of processors seen by the OS or Java runtime
@see docs.oracle.com/javase/6/docs/api/java/lang/Runtime.html#availableProcessors()
# File lib/concurrent-ruby/concurrent/utility/processor_counter.rb, line 86 def self.processor_count processor_counter.processor_count end
Use logger created by create_simple_logger to log concurrent-ruby messages.
# File lib/concurrent-ruby/concurrent/concern/logging.rb, line 63 def self.use_simple_logger(level = Logger::FATAL, output = $stderr) Concurrent.global_logger = create_simple_logger level, output end
Use logger created by create_stdlib_logger to log concurrent-ruby messages. @deprecated
# File lib/concurrent-ruby/concurrent/concern/logging.rb, line 96 def self.use_stdlib_logger(level = Logger::FATAL, output = $stderr) Concurrent.global_logger = create_stdlib_logger level, output end
Private Instance Methods
Abort a currently running transaction - see ‘Concurrent::atomically`.
# File lib/concurrent-ruby/concurrent/tvar.rb, line 139 def abort_transaction raise Transaction::AbortError.new end
Run a block that reads and writes ‘TVar`s as a single atomic transaction. With respect to the value of `TVar` objects, the transaction is atomic, in that it either happens or it does not, consistent, in that the `TVar` objects involved will never enter an illegal state, and isolated, in that transactions never interfere with each other. You may recognise these properties from database transactions.
There are some very important and unusual semantics that you must be aware of:
-
Most importantly, the block that you pass to atomically may be executed
more than once. In most cases your code should be free of side-effects, except for via TVar.
-
If an exception escapes an atomically block it will abort the transaction.
-
It is undefined behaviour to use callcc or Fiber with atomically.
-
If you create a new thread within an atomically, it will not be part of
the transaction. Creating a thread counts as a side-effect.
Transactions within transactions are flattened to a single transaction.
@example
a = new TVar(100_000) b = new TVar(100) Concurrent::atomically do a.value -= 10 b.value += 10 end
# File lib/concurrent-ruby/concurrent/tvar.rb, line 82 def atomically raise ArgumentError.new('no block given') unless block_given? # Get the current transaction transaction = Transaction::current # Are we not already in a transaction (not nested)? if transaction.nil? # New transaction begin # Retry loop loop do # Create a new transaction transaction = Transaction.new Transaction::current = transaction # Run the block, aborting on exceptions begin result = yield rescue Transaction::AbortError => e transaction.abort result = Transaction::ABORTED rescue Transaction::LeaveError => e transaction.abort break result rescue => e transaction.abort raise e end # If we can commit, break out of the loop if result != Transaction::ABORTED if transaction.commit break result end end end ensure # Clear the current transaction Transaction::current = nil end else # Nested transaction - flatten it and just run the block yield end end
# File lib/concurrent-ruby/concurrent/dataflow.rb, line 56 def call_dataflow(method, executor, *inputs, &block) raise ArgumentError.new('an executor must be provided') if executor.nil? raise ArgumentError.new('no block given') unless block_given? unless inputs.all? { |input| input.is_a? IVar } raise ArgumentError.new("Not all dependencies are IVars.\nDependencies: #{ inputs.inspect }") end result = Future.new(executor: executor) do values = inputs.map { |input| input.send(method) } block.call(*values) end if inputs.empty? result.execute else counter = DependencyCounter.new(inputs.size) { result.execute } inputs.each do |input| input.add_observer counter end end result end
Dataflow allows you to create a task that will be scheduled when all of its data dependencies are available. {include:file:docs-source/dataflow.md}
@param [Future] inputs zero or more ‘Future` operations that this dataflow depends upon
@yield The operation to perform once all the dependencies are met @yieldparam [Future] inputs each of the ‘Future` inputs to the dataflow @yieldreturn [Object] the result of the block operation
@return [Object] the result of all the operations
@raise [ArgumentError] if no block is given @raise [ArgumentError] if any of the inputs are not ‘IVar`s
# File lib/concurrent-ruby/concurrent/dataflow.rb, line 34 def dataflow(*inputs, &block) dataflow_with(Concurrent.global_io_executor, *inputs, &block) end
# File lib/concurrent-ruby/concurrent/dataflow.rb, line 44 def dataflow!(*inputs, &block) dataflow_with!(Concurrent.global_io_executor, *inputs, &block) end
# File lib/concurrent-ruby/concurrent/dataflow.rb, line 39 def dataflow_with(executor, *inputs, &block) call_dataflow(:value, executor, *inputs, &block) end
# File lib/concurrent-ruby/concurrent/dataflow.rb, line 49 def dataflow_with!(executor, *inputs, &block) call_dataflow(:value!, executor, *inputs, &block) end
Leave a transaction without committing or aborting - see ‘Concurrent::atomically`.
# File lib/concurrent-ruby/concurrent/tvar.rb, line 144 def leave_transaction raise Transaction::LeaveError.new end
@!macro monotonic_get_time
Returns the current time as tracked by the application monotonic clock. @param [Symbol] unit the time unit to be returned, can be either :float_second, :float_millisecond, :float_microsecond, :second, :millisecond, :microsecond, or :nanosecond default to :float_second. @return [Float] The current monotonic time since some unspecified starting point @!macro monotonic_clock_warning
# File lib/concurrent-ruby/concurrent/utility/monotonic_time.rb, line 15 def monotonic_time(unit = :float_second) Process.clock_gettime(Process::CLOCK_MONOTONIC, unit) end