by having Mailbox extend ForkJoinTask and using the fact that ForkJoinTask.exec when returning
false does not set completion on the task, so it is free to be resubmitted to the ForkJoinPool
without reinitialization.
Also adds the ability to use fork() when the currentThread is a worker thread of the pool that we want to execute on.
Adds a JMH benchmark for both the ping-pong performance and pipelined throughput.
Conflicts:
project/AkkaBuild.scala
* The filter is used by the LoggingAdapter before publishing
to the event bus
* Slf4jLoggingFilter uses backend log level configuration
(e.g. logback.xml)
- add “mailbox-requirement” key to dispatcher section
- split out mailbox section, add akka.actor.default-mailbox
- rewrite findMarker method and use it for Props.create() and getting
the required mailbox of an actor
- add ProducesMessageQueue trait for MailboxType so that requirements
can be checked before trying to create the actor for real
- verify actor as well as dispatcher requirements for message queue
before creation, even in remote-deployed case
- change MessageDispatcher constructor to take a Configurator, add that
to migration guide
- remove some outdated ones (which were already fixed but not removed)
- make socket options fully configurable in NettyTransport
- also apply socket options to outbound connections, not only the listen
socket
- remove timeout from taking the lock in RepointableActorRef, because we
cannot afford to lose system messages anymore, so we should rather
block the sender
- Moved system messages to their own package.
- All queueing operations are now hidden behind a SystemMessageList value class
- Introduced dual SystemMessageList types to encode the ordering in the type.
- Protects against accidentally missed reverse calls or accidentally reversed lists
- Makes ordering expectations by fields/parameters explicit
- Fixed serialization tests
- Fixes to logging in HierarchyStressSpec
* Sending to a previous incarnation of an actor shall fail,
to make remote actors work the same way as local ones (in
the sense that after Terminated() the ref is not working anymore)
* Changed equality of ActorRef to take the uid into account
* Parse uid fragment in RelativeActorPath and ActorPathExtractor
* Handle uid in getChild and in RemoteSystemDaemon
* Use toSerializationFormat and toSerializationFormatWithAddress
in serialization
* Replaced var uid in ActorCell and ChildRestartStats with
constructor parameters (path)
* Create the uid in one single place, in makeChild in parent
* Handle ActorRef with and without uid in DeathWatch
* Optimize ActorPath.toString and friends
* Update documentation and migration guide
* Make SystemMessage extend Serializable to avoid ambiguity when
setting serialization-bindings.
* Set serialVersionUID in SystemMessages and create tests to
ensure binary formats remain unchanged.
* Add tests for reference.conf's serialization settings.
* Make some existing serialization tests more robust.
Removed boilerplate from serialization tests
Use actual reference.conf; tidy up
Make serialization compatible
* When a def starts with if and is not a oneliner the if
should be on a new line.
* The reason is that it might be easy to miss the if when
reading the code.
The MonitorableThreadFactory.copy method does not take into account the
counter:AtomicLong field, which then always starts out at zero for every
new dispatcher; the PinnedDispatcher creates one dispatcher per actor,
hence resulting in identical thread names if used by multiple actors.
Solution: add the counter to the case class’ arguments
- parent just checked for system messages and finds none
- Supervise is enqueued to parent
- Failed is enqueued to the parent
- Failed is processed before Supervise and thus dropped due to
non-matching UID
remedy is to enqueue NullMessage after enqueuing Supervise but before
really starting the actor, so that if NullMessage overtakes Supervise
the loop in processMailbox will then pick up Supervise next
- every ActorCell has a UID which is generated from ThreadLocalRandom
and also sent with Supervise and Failed messages, and the supervisor
simply does not act upon failures from children whose UID does not
match
This has brought to light some interesting effects (aka bugs) both in
the general implementation as well as in previous fixes.
SupervisorHierarchySpec is without TODOs now and GREEN.
