Class: Concurrent::Promises::Channel
Relationships & Source Files | |
Super Chains via Extension / Inclusion / Inheritance | |
Class Chain:
|
|
Instance Chain:
|
|
Inherits: |
Concurrent::Synchronization::Object
|
Defined in: | lib/concurrent-ruby-edge/concurrent/edge/channel.rb |
Overview
**Edge Features** are under active development and may change frequently.
-
Deprecations are not added before incompatible changes.
-
Edge
version: major is always 0, minor bump means incompatible change, patch bump means compatible change. -
Edge
features may also lack tests and documentation. -
Features developed in
concurrent-ruby-edge
are expected to move toconcurrent-ruby
when finalised.
A first in first out channel that accepts messages with push family of methods and returns messages with pop family of methods. Pop and push operations can be represented as futures, see #pop_op and #push_op. The capacity of the channel can be limited to support back pressure, use capacity option in Concurrent#initialize. #pop method blocks ans #pop_op returns pending future if there is no message in the channel. If the capacity is limited the #push method blocks and #push_op returns pending future.
==== Examples
Let's start by creating a channel with a capacity of 2 messages.
ch = {Concurrent::Promises::Channel.new} 2
=== => #<Concurrent::Promises::Channel:0x000002 capacity taken 0 of 2>
We push 3 messages, then it can be observed that the last thread pushing is sleeping since the channel is full.
threads = {Array.new}(3) { |i| {Thread.new} { ch.push message: i } }
sleep 0.01 # let the threads run
threads
=== => [#<Thread:0x000003@channel.in.md:14 dead>,
=== #<Thread:0x000004@channel.in.md:14 dead>,
=== #<Thread:0x000005@channel.in.md:14 sleep_forever>]
When message is popped the last thread continues and finishes as well.
ch.pop # => {:message=>1}
threads.map(&:join)
=== => [#<Thread:0x000003@channel.in.md:14 dead>,
=== #<Thread:0x000004@channel.in.md:14 dead>,
=== #<Thread:0x000005@channel.in.md:14 dead>]
Same principle applies to popping as well. There are now 2 messages int he channel. Lets create 3 threads trying to pop a message, one will be blocked until new messages is pushed.
threads = {Array.new}(3) { |i| {Thread.new} { ch.pop } }
sleep 0.01 # let the threads run
threads
=== => [#<Thread:0x000006@channel.in.md:32 dead>,
=== #<Thread:0x000007@channel.in.md:32 dead>,
=== #<Thread:0x000008@channel.in.md:32 sleep_forever>]
ch.push message: 3
=== => #<Concurrent::Promises::Channel:0x000002 capacity taken 0 of 2>
threads.map(&:value)
=== => [{:=>0}, {:=>2}, {:=>3}]
===== Promises integration
However this channel is implemented to integrate with promises therefore all operations can be represented as futures.
ch = {Concurrent::Promises::Channel.new} 2
=== => #<Concurrent::Promises::Channel:0x000009 capacity taken 0 of 2>
push_operations = {Array.new}(3) { |i| ch.push_op message: i }
=== => [#<Concurrent::Promises::Future:0x00000a fulfilled with #<Concurrent::Promises::Channel:0x000009 capacity taken 2 of 2>>,
=== #<Concurrent::Promises::Future:0x00000b fulfilled with #<Concurrent::Promises::Channel:0x000009 capacity taken 2 of 2>>,
=== #<Concurrent::Promises::ResolvableFuture:0x00000c pending>]
We do not have to sleep here letting the futures execute as Threads. Since there is capacity for 2 messages the Promises are immediately resolved without ever allocating a Thread to execute. Push and pop operations are often more efficient. The remaining pending push operation will also never require another thread, instead it will resolve when a message is popped from the channel making a space for a new message.
ch.pop_op.value! # => {:message=>0}
push_operations.map(&:value!)
=== => [#<Concurrent::Promises::Channel:0x000009 capacity taken 2 of 2>,
=== #<Concurrent::Promises::Channel:0x000009 capacity taken 2 of 2>,
=== #<Concurrent::Promises::Channel:0x000009 capacity taken 2 of 2>]
pop_operations = {Array.new}(3) { |i| ch.pop_op }
=== => [#<Concurrent::Promises::ResolvableFuture:0x00000d fulfilled with {:message=>1}>,
=== #<Concurrent::Promises::ResolvableFuture:0x00000e fulfilled with {:message=>2}>,
=== #<Concurrent::Promises::ResolvableFuture:0x00000f pending>]
ch.push message: 3 # (push|pop) can be freely mixed with (push_o|pop_op)
pop_operations.map(&:value)
=== => [{:=>1}, {:=>2}, {:=>3}]
===== Selecting over channels
A selection over channels can be created with the .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.
ch1 = {Concurrent::Promises::Channel.new} 2
=== => #<Concurrent::Promises::Channel:0x000010 capacity taken 0 of 2>
ch2 = {Concurrent::Promises::Channel.new} 2
=== => #<Concurrent::Promises::Channel:0x000011 capacity taken 0 of 2>
ch1.push 1
=== => #<Concurrent::Promises::Channel:0x000010 capacity taken 1 of 2>
ch2.push 2
=== => #<Concurrent::Promises::Channel:0x000011 capacity taken 1 of 2>
<code>Channel</code>.select([ch1, ch2])
=== => [#<Concurrent::Promises::Channel:0x000010 capacity taken 0 of 2>, 1]
ch1.select(ch2)
=== => [#<Concurrent::Promises::Channel:0x000011 capacity taken 0 of 2>, 2]
{Concurrent::Promises.future} { 3 + 4 }.then_channel_push(ch1)
=== => #<Concurrent::Promises::Future:0x000012 pending>
<code>Channel</code>.
# or `ch1.select_op(ch2)` would be equivalent
select_op([ch1, ch2]).
then('got number %03d from ch%d') { |(channel, value), format|
format format, value, [ch1, ch2].index(channel).succ
}.value! # => "got number 007 from ch1"
===== try_
variants
All blocking operations (#pop, #push, #select) have non-blocking variant
with try_
prefix.
They always return immediately and indicate either success or failure.
ch
=== => #<Concurrent::Promises::Channel:0x000009 capacity taken 0 of 2>
ch.try_push 1 # => true
ch.try_push 2 # => true
ch.try_push 3 # => false
ch.try_pop # => 1
ch.try_pop # => 2
ch.try_pop # => nil
===== Timeouts
All blocking operations (#pop, #push, #select) have a timeout option.
Similar to try_
variants it will indicate success or timing out,
when the timeout option is used.
ch
=== => #<Concurrent::Promises::Channel:0x000009 capacity taken 0 of 2>
ch.push 1, 0.01 # => true
ch.push 2, 0.01 # => true
ch.push 3, 0.01 # => false
ch.pop 0.01 # => 1
ch.pop 0.01 # => 2
ch.pop 0.01 # => nil
===== Backpressure
Most importantly the channel can be used to create systems with backpressure. A self adjusting system where the producers will slow down if the consumers are not keeping up.
channel = {Concurrent::Promises::Channel.new} 2
=== => #<Concurrent::Promises::Channel:0x000013 capacity taken 0 of 2>
log = {Concurrent::Array.new} # => []
producers = {Array.new} 2 do |i|
Thread.new(i) do |i|
4.times do |j|
log.push format "producer %d pushing %d", i, j
channel.push [i, j]
end
end
end
=== => [#<Thread:0x000014@channel.in.md:133 run>,
=== #<Thread:0x000015@channel.in.md:133 run>]
consumers = {Array.new} 4 do |i|
Thread.new(i) do |consumer|
2.times do |j|
from, = channel.pop
log.push format "consumer %d got %d. payload %d from producer %d",
consumer, j, , from
do_stuff
end
end
end
=== => [#<Thread:0x000016@channel.in.md:142 run>,
=== #<Thread:0x000017@channel.in.md:142 run>,
=== #<Thread:0x000018@channel.in.md:142 run>,
=== #<Thread:0x000019@channel.in.md:142 run>]
=== wait for all to finish
producers.map(&:join)
=== => [#<Thread:0x000014@channel.in.md:133 dead>,
=== #<Thread:0x000015@channel.in.md:133 dead>]
consumers.map(&:join)
=== => [#<Thread:0x000016@channel.in.md:142 dead>,
=== #<Thread:0x000017@channel.in.md:142 dead>,
=== #<Thread:0x000018@channel.in.md:142 dead>,
=== #<Thread:0x000019@channel.in.md:142 dead>]
=== investigate log
log
=== => ["producer 0 pushing 0",
=== "producer 0 pushing 1",
=== "producer 0 pushing 2",
=== "producer 1 pushing 0",
=== "consumer 0 got 0. payload 0 from producer 0",
=== "producer 0 pushing 3",
=== "consumer 1 got 0. payload 1 from producer 0",
=== "consumer 2 got 0. payload 2 from producer 0",
=== "consumer 3 got 0. payload 0 from producer 1",
=== "producer 1 pushing 1",
=== "producer 1 pushing 2",
=== "consumer 1 got 1. payload 3 from producer 0",
=== "producer 1 pushing 3",
=== "consumer 2 got 1. payload 1 from producer 1",
=== "consumer 3 got 1. payload 2 from producer 1",
=== "consumer 0 got 1. payload 3 from producer 1"]
The producers are much faster than consumers
(since they do_stuff
which takes some time)
but as it can be seen from the log they fill the channel
and then they slow down
until there is space available in the channel.
If permanent allocation of threads to the producers and consumers has to be avoided, the threads can be replaced with promises that run a thread pool.
channel = {Concurrent::Promises::Channel.new} 2
=== => #<Concurrent::Promises::Channel:0x00001a capacity taken 0 of 2>
log = {Concurrent::Array.new} # => []
def produce(channel, log, producer, i)
log.push format "producer %d pushing %d", producer, i
channel.push_op([producer, i]).then do
i + 1 < 4 ? produce(channel, log, producer, i + 1) : :done
end
end # => :produce
def consume(channel, log, consumer, i)
channel.pop_op.then(consumer, i) do |(from, ), consumer, i|
log.push format "consumer %d got %d. payload %d from producer %d",
consumer, i, , from
do_stuff
i + 1 < 2 ? consume(channel, log, consumer, i + 1) : :done
end
end # => :consume
producers = {Array.new} 2 do |i|
Concurrent::Promises.future(channel, log, i) { |*args| produce *args, 0 }.run
end
=== => [#<Concurrent::Promises::Future:0x00001b pending>,
=== #<Concurrent::Promises::Future:0x00001c pending>]
consumers = {Array.new} 4 do |i|
Concurrent::Promises.future(channel, log, i) { |*args| consume *args, 0 }.run
end
=== => [#<Concurrent::Promises::Future:0x00001d pending>,
=== #<Concurrent::Promises::Future:0x00001e pending>,
=== #<Concurrent::Promises::Future:0x00001f pending>,
=== #<Concurrent::Promises::Future:0x000020 pending>]
=== wait for all to finish
producers.map(&:value!) # => [:done, :done]
consumers.map(&:value!) # => [:done, :done, :done, :done]
=== investigate log
log
=== => ["producer 0 pushing 0",
=== "producer 1 pushing 0",
=== "producer 1 pushing 1",
=== "consumer 1 got 0. payload 0 from producer 1",
=== "consumer 2 got 0. payload 1 from producer 1",
=== "producer 0 pushing 1",
=== "producer 0 pushing 2",
=== "producer 0 pushing 3",
=== "producer 1 pushing 2",
=== "consumer 0 got 0. payload 0 from producer 0",
=== "consumer 3 got 0. payload 1 from producer 0",
=== "producer 1 pushing 3",
=== "consumer 2 got 1. payload 2 from producer 0",
=== "consumer 1 got 1. payload 3 from producer 0",
=== "consumer 3 got 1. payload 3 from producer 1",
=== "consumer 0 got 1. payload 2 from producer 1"]
===== Synchronization of workers by passing a value
If the capacity of the channel is zero then any push operation will succeed only when there is a matching pop operation which can take the message. The operations have to be paired to succeed.
channel = {Concurrent::Promises::Channel.new} 0
=== => #<Concurrent::Promises::Channel:0x000021 capacity taken 0 of 0>
thread = {Thread.new} { channel.pop }; sleep 0.01
=== allow the thread to go to sleep
thread
=== => #<Thread:0x000022@channel.in.md:214 sleep_forever>
=== succeeds because there is matching pop operation waiting in the thread
channel.try_push(:v1) # => true
=== remains pending, since there is no matching operation
push = channel.push_op(:v2)
=== => #<Concurrent::Promises::ResolvableFuture:0x000023 pending>
thread.value # => :v1
=== the push operation resolves as a pairing pop is called
channel.pop # => :v2
push
=== => #<Concurrent::Promises::ResolvableFuture:0x000023 fulfilled with #<Concurrent::Promises::Channel:0x000021 capacity taken 0 of 0>>
Constant Summary
-
ANY =
An object which matches anything (with
#===
)Object.new.tap do |any| def any.===(other) true end def any.to_s 'ANY' end end
-
NOTHING =
private
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 48Object.new
-
UNLIMITED_CAPACITY =
Default capacity of the
Channel
, makes it accept unlimited number of messages.::Object.new
Class Attribute Summary
Synchronization::Object
- Inherited
Class Method Summary
-
.new(capacity = UNLIMITED_CAPACITY) ⇒ Channel
constructor
Create channel.
- .select(channels, timeout = nil) ⇒ ::Array(Channel, Object)?
- .select_matching(matcher, channels, timeout = nil) ⇒ ::Array(Channel, Object)?
- .select_op(channels, probe = Promises.resolvable_future) ⇒ Future(::Array(Channel, Object))
- .select_op_matching(matcher, channels, probe = Promises.resolvable_future) ⇒ Future(::Array(Channel, Object))
- .try_select(channels) ⇒ ::Array(Channel, Object)
- .try_select_matching(matcher, channels) ⇒ ::Array(Channel, Object)
Synchronization::Object
- Inherited
.atomic_attribute?, .atomic_attributes, | |
.attr_atomic | Creates methods for reading and writing to a instance variable with volatile (Java) semantic as |
.attr_volatile | Creates methods for reading and writing (as |
.ensure_safe_initialization_when_final_fields_are_present | For testing purposes, quite slow. |
.new | Has to be called by children. |
.safe_initialization!, .define_initialize_atomic_fields |
Synchronization::AbstractObject
- Inherited
Instance Method Summary
- #capacity ⇒ Integer
-
#inspect
Alias for #to_s.
-
#peek(no_value = nil) ⇒ Object, no_value
Behaves as #try_pop but it does not remove the message from the channel.
-
#peek_matching(matcher, no_value = nil) ⇒ Object, no_value
Behaves as #try_pop but it does not remove the message from the channel.
-
#pop(timeout = nil, timeout_value = nil) ⇒ Object?
Blocks current thread until a message is available in the channel for popping.
-
#pop_matching(matcher, timeout = nil, timeout_value = nil) ⇒ Object?
Blocks current thread until a message is available in the channel for popping.
-
#pop_op(probe = Promises.resolvable_future) ⇒ Future(Object)
Returns a future witch will become fulfilled with a value from the channel when one is available.
-
#pop_op_matching(matcher, probe = Promises.resolvable_future) ⇒ Future(Object)
Returns a future witch will become fulfilled with a value from the channel when one is available.
-
#push(message, timeout = nil) ⇒ self, ...
Blocks current thread until the message is pushed into the channel.
-
#push_op(message) ⇒ ResolvableFuture(self)
Returns future which will fulfill when the message is pushed to the channel.
-
#select(channels, timeout = nil) ⇒ ::Array(Channel, Object)?
As #select_op but does not return future, it block current thread instead until there is a message available in the receiver or in any of the channels.
-
#select_matching(matcher, channels, timeout = nil) ⇒ ::Array(Channel, Object)?
As #select_op but does not return future, it block current thread instead until there is a message available in the receiver or in any of the channels.
-
#select_op(channels, probe = Promises.resolvable_future) ⇒ ResolvableFuture(::Array(Channel, Object))
When message is available in the receiver or any of the provided channels the future is fulfilled with a channel message pair.
-
#select_op_matching(matcher, channels, probe = Promises.resolvable_future) ⇒ ResolvableFuture(::Array(Channel, Object))
When message is available in the receiver or any of the provided channels the future is fulfilled with a channel message pair.
- #size ⇒ Integer
- #to_s ⇒ String (also: #inspect)
-
#try_pop(no_value = nil) ⇒ Object, no_value
Pop a message from the channel if there is one available.
-
#try_pop_matching(matcher, no_value = nil) ⇒ Object, no_value
Pop a message from the channel if there is one available.
-
#try_push(message) ⇒ true, false
Push the message into the channel if there is space available.
-
#try_select(channels) ⇒ ::Array(Channel, Object)?
If message is available in the receiver or any of the provided channels the channel message pair is returned.
-
#try_select_matching(matcher, channels) ⇒ ::Array(Channel, Object)?
If message is available in the receiver or any of the provided channels the channel message pair is returned.
- #ns_consume_pending_push(matcher, remove = true) private
- #ns_pop_op(matcher, probe, include_channel) private
- #ns_shift_message(matcher, remove = true) private
- #ns_try_push(message) private
- #partial_select_op(matcher, probe) private
Synchronization::Object
- Inherited
Synchronization::Volatile
- Included
Synchronization::AbstractObject
- Inherited
Constructor Details
.new(capacity = UNLIMITED_CAPACITY) ⇒ Channel
Create channel.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 64
def initialize(capacity = UNLIMITED_CAPACITY) super() @Capacity = capacity @Mutex = Mutex.new # TODO (pitr-ch 28-Jan-2019): consider linked lists or other data structures for following attributes, things are being deleted from the middle @Probes = [] @Messages = [] @PendingPush = [] end
Class Method Details
.select(channels, timeout = nil) ⇒ ::Array
(Channel
, Object)?
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 322
def select(channels, timeout = nil) channels.first.select(channels[1..-1], timeout) end
.select_matching(matcher, channels, timeout = nil) ⇒ ::Array
(Channel
, Object)?
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 340
def select_matching(matcher, channels, timeout = nil) channels.first.select_matching(matcher, channels[1..-1], timeout) end
.select_op(channels, probe = Promises.resolvable_future) ⇒ Future(::Array
(Channel
, Object))
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 316
def select_op(channels, probe = Promises.resolvable_future) channels.first.select_op(channels[1..-1], probe) end
.select_op_matching(matcher, channels, probe = Promises.resolvable_future) ⇒ Future(::Array
(Channel
, Object))
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 334
def select_op_matching(matcher, channels, probe = Promises.resolvable_future) channels.first.select_op_matching(matcher, channels[1..-1], probe) end
.try_select(channels) ⇒ ::Array
(Channel
, Object)
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 310
def try_select(channels) channels.first.try_select(channels[1..-1]) end
.try_select_matching(matcher, channels) ⇒ ::Array
(Channel
, Object)
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 328
def try_select_matching(matcher, channels) channels.first.try_select_matching(matcher, channels[1..-1]) end
Instance Method Details
#capacity ⇒ Integer
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 295
def capacity @Capacity end
#inspect
Alias for #to_s.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 304
alias_method :inspect, :to_s
#ns_consume_pending_push(matcher, remove = true) (private)
[ GitHub ]# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 397
def ns_consume_pending_push(matcher, remove = true) i = 0 while true , pushed = @PendingPush[i, 2] return NOTHING unless pushed if matcher === resolved = pushed.resolved? @PendingPush[i, 2] = [] if remove || resolved # can fail if timed-out, so try without error if remove ? pushed.fulfill(self, false) : !resolved # pushed fulfilled so actually push the message return end end i += 2 end end
#ns_pop_op(matcher, probe, include_channel) (private)
[ GitHub ]# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 352
def ns_pop_op(matcher, probe, include_channel) = matcher # got message from buffer if != NOTHING if probe.fulfill(include_channel ? [self, ] : , false) = ns_consume_pending_push ANY @Messages.push unless == NOTHING else @Messages.unshift end return probe end # no message in buffer, try to pair with a pending push i = 0 while true , pushed = @PendingPush[i, 2] break if pushed.nil? if matcher === value = include_channel ? [self, ] : if Promises::Resolvable.atomic_resolution(probe => [true, value, nil], pushed => [true, self, nil]) @PendingPush[i, 2] = [] return probe end if probe.resolved? return probe end # so pushed.resolved? has to be true, remove the push @PendingPush[i, 2] = [] end i += 2 end # no push to pair with # TODO (pitr-ch 11-Jan-2019): clear up probes when timed out, use callback @Probes.push probe, include_channel, matcher if probe.pending? return probe end
#ns_shift_message(matcher, remove = true) (private)
[ GitHub ]#ns_try_push(message) (private)
[ GitHub ]# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 417
def ns_try_push( ) i = 0 while true probe, include_channel, matcher = @Probes[i, 3] break unless probe if matcher === && probe.fulfill(include_channel ? [self, ] : , false) @Probes[i, 3] = [] return true end i += 3 end if @Capacity > @Messages.size @Messages.push true else false end end
#partial_select_op(matcher, probe) (private)
[ GitHub ]# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 346
def partial_select_op(matcher, probe) @Mutex.synchronize { ns_pop_op(matcher, probe, true) } end
#peek(no_value = nil) ⇒ Object, no_value
Behaves as #try_pop but it does not remove the message from the channel
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 209
def peek(no_value = nil) peek_matching ANY, no_value end
#peek_matching(matcher, no_value = nil) ⇒ Object, no_value
Behaves as #try_pop but it does not remove the message from the channel
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 215
def peek_matching(matcher, no_value = nil) @Mutex.synchronize do = matcher, false return if != NOTHING = ns_consume_pending_push matcher, false return != NOTHING ? : no_value end end
#pop(timeout = nil, timeout_value = nil) ⇒ Object?
This function potentially blocks current thread until it can continue. Be careful it can deadlock.
Blocks current thread until a message is available in the channel for popping.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 177
def pop(timeout = nil, timeout_value = nil) pop_matching ANY, timeout, timeout_value end
#pop_matching(matcher, timeout = nil, timeout_value = nil) ⇒ Object?
This function potentially blocks current thread until it can continue. Be careful it can deadlock.
Blocks current thread until a message is available in the channel for popping.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 183
def pop_matching(matcher, timeout = nil, timeout_value = nil) # TODO (pitr-ch 27-Jan-2019): should it try to match pending pushes if it fails to match in the buffer? Maybe only if the size is zero. It could be surprising if it's used as a throttle it might be expected that it will not pop if buffer is full of messages which di not match, it might it expected it will block until the message is added to the buffer # that it returns even if the buffer is full. User might expect that it has to be in the buffer first. probe = @Mutex.synchronize do = matcher if == NOTHING = ns_consume_pending_push matcher return if != NOTHING else = ns_consume_pending_push ANY @Messages.push unless == NOTHING return end probe = Promises.resolvable_future @Probes.push probe, false, matcher probe end probe.value!(timeout, timeout_value, [true, timeout_value, nil]) end
#pop_op(probe = Promises.resolvable_future) ⇒ Future(Object)
Returns a future witch will become fulfilled with a value from the channel when one is available. If it is later waited on the operation with a timeout e.g.‘channel.pop_op.wait(1)` it will not prevent the channel to fulfill the operation later after the timeout. The operation has to be either processed later “`ruby pop_op = channel.pop_op if pop_op.wait(1)
pop_op.value
else
pop_op.then { || }
end “‘ or the operation can be prevented from completion after timing out by using channel.pop_op.wait(1, [true, nil, nil])
. It will fulfill the operation on timeout preventing channel from doing the operation, e.g. popping a message.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 160
def pop_op(probe = Promises.resolvable_future) @Mutex.synchronize { ns_pop_op(ANY, probe, false) } end
#pop_op_matching(matcher, probe = Promises.resolvable_future) ⇒ Future(Object)
Returns a future witch will become fulfilled with a value from the channel when one is available. If it is later waited on the operation with a timeout e.g.‘channel.pop_op.wait(1)` it will not prevent the channel to fulfill the operation later after the timeout. The operation has to be either processed later “`ruby pop_op = channel.pop_op if pop_op.wait(1)
pop_op.value
else
pop_op.then { || }
end “‘ or the operation can be prevented from completion after timing out by using channel.pop_op.wait(1, [true, nil, nil])
. It will fulfill the operation on timeout preventing channel from doing the operation, e.g. popping a message.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 166
def pop_op_matching(matcher, probe = Promises.resolvable_future) @Mutex.synchronize { ns_pop_op(matcher, probe, false) } end
#push(message, timeout = nil) ⇒ self
, ...
This function potentially blocks current thread until it can continue. Be careful it can deadlock.
Blocks current thread until the message is pushed into the channel.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 120
def push(, timeout = nil) pushed_op = @Mutex.synchronize do return timeout ? true : self if ns_try_push( ) pushed = Promises.resolvable_future # TODO (pitr-ch 06-Jan-2019): clear timed out pushes in @PendingPush, null messages @PendingPush.push , pushed pushed end result = pushed_op.wait!(timeout, [true, self, nil]) result == pushed_op ? self : result end
#push_op(message) ⇒ ResolvableFuture(self
)
Returns future which will fulfill when the message is pushed to the channel. If it is later waited on the operation with a timeout e.g.‘channel.pop_op.wait(1)` it will not prevent the channel to fulfill the operation later after the timeout. The operation has to be either processed later “`ruby pop_op = channel.pop_op if pop_op.wait(1)
pop_op.value
else
pop_op.then { || }
end “‘ or the operation can be prevented from completion after timing out by using channel.pop_op.wait(1, [true, nil, nil])
. It will fulfill the operation on timeout preventing channel from doing the operation, e.g. popping a message.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 101
def push_op( ) @Mutex.synchronize do if ns_try_push( ) Promises.fulfilled_future self else pushed = Promises.resolvable_future @PendingPush.push , pushed return pushed end end end
#select(channels, timeout = nil) ⇒ ::Array
(Channel
, Object)?
This function potentially blocks current thread until it can continue. Be careful it can deadlock.
As #select_op but does not return future, it block current thread instead until there is a message available in the receiver or in any of the channels.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 278
def select(channels, timeout = nil) select_matching ANY, channels, timeout end
#select_matching(matcher, channels, timeout = nil) ⇒ ::Array
(Channel
, Object)?
This function potentially blocks current thread until it can continue. Be careful it can deadlock.
As #select_op but does not return future, it block current thread instead until there is a message available in the receiver or in any of the channels.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 284
def select_matching(matcher, channels, timeout = nil) probe = select_op_matching(matcher, channels) probe.value!(timeout, nil, [true, nil, nil]) end
#select_op(channels, probe = Promises.resolvable_future) ⇒ ResolvableFuture(::Array
(Channel
, Object))
When message is available in the receiver or any of the provided channels the future is fulfilled with a channel message pair. The returned channel is the origin of the message. If it is later waited on the operation with a timeout e.g.‘channel.pop_op.wait(1)` it will not prevent the channel to fulfill the operation later after the timeout. The operation has to be either processed later “`ruby pop_op = channel.pop_op if pop_op.wait(1)
pop_op.value
else
pop_op.then { || }
end “‘ or the operation can be prevented from completion after timing out by using channel.pop_op.wait(1, [true, nil, nil])
. It will fulfill the operation on timeout preventing channel from doing the operation, e.g. popping a message.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 257
def select_op(channels, probe = Promises.resolvable_future) select_op_matching ANY, channels, probe end
#select_op_matching(matcher, channels, probe = Promises.resolvable_future) ⇒ ResolvableFuture(::Array
(Channel
, Object))
When message is available in the receiver or any of the provided channels the future is fulfilled with a channel message pair. The returned channel is the origin of the message. If it is later waited on the operation with a timeout e.g.‘channel.pop_op.wait(1)` it will not prevent the channel to fulfill the operation later after the timeout. The operation has to be either processed later “`ruby pop_op = channel.pop_op if pop_op.wait(1)
pop_op.value
else
pop_op.then { || }
end “‘ or the operation can be prevented from completion after timing out by using channel.pop_op.wait(1, [true, nil, nil])
. It will fulfill the operation on timeout preventing channel from doing the operation, e.g. popping a message.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 263
def select_op_matching(matcher, channels, probe = Promises.resolvable_future) [self, *channels].each { |ch| ch.partial_select_op matcher, probe } probe end
#size ⇒ Integer
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 290
def size @Mutex.synchronize { @Messages.size } end
#to_s ⇒ String
Also known as: #inspect
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 300
def to_s format '%s capacity taken %s of %s>', super[0..-2], size, @Capacity end
#try_pop(no_value = nil) ⇒ Object, no_value
Pop a message from the channel if there is one available.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 138
def try_pop(no_value = nil) try_pop_matching ANY, no_value end
#try_pop_matching(matcher, no_value = nil) ⇒ Object, no_value
Pop a message from the channel if there is one available.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 145
def try_pop_matching(matcher, no_value = nil) @Mutex.synchronize do = matcher return if != NOTHING = ns_consume_pending_push matcher return != NOTHING ? : no_value end end
#try_push(message) ⇒ true
, false
Push the message into the channel if there is space available.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 77
def try_push( ) @Mutex.synchronize { ns_try_push( ) } end
#try_select(channels) ⇒ ::Array
(Channel
, Object)?
If message is available in the receiver or any of the provided channels the channel message pair is returned. If there is no message nil is returned. The returned channel is the origin of the message.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 232
def try_select(channels) try_select_matching ANY, channels end
#try_select_matching(matcher, channels) ⇒ ::Array
(Channel
, Object)?
If message is available in the receiver or any of the provided channels the channel message pair is returned. If there is no message nil is returned. The returned channel is the origin of the message.
# File 'lib/concurrent-ruby-edge/concurrent/edge/channel.rb', line 238
def try_select_matching(matcher, channels) = nil channel = [self, *channels].find do |ch| = ch.try_pop_matching(matcher, NOTHING) != NOTHING end channel ? [channel, ] : nil end