| Age | Commit message (Collapse) | Author |
|---|
|
This change adds helper methods for simplifying the request-response
pattern commonly used in models.
|
|
This change fixes the issue where messages are not delivered in FIFO
order due to the internal priority not guaranteeing insertion order. For
now, we fix this issue by adding a unique increasing identifier to each
event in the queue.
|
|
Add support for Gitlab CI
See merge request opendc/opendc-simulator!22
|
|
This change adds the configuration of Gitlab CI in order to test on our
internal Gitlab instance.
|
|
This change adds experimental support for selecting on ports. This
allows the user to receive messages from multiple channels at the same
time.
|
|
This change adds a prototype implementation of the revised version of
the API of version 2.0 of the simulator.
|
|
This change introduces the revised API design for version 2.0 of the
OpenDC simulator. This version drops built-in support for Java and
instead opts to build on Kotlin coroutines to simplify the API surface.
During development of and experimentation with the previous API for version
2.x, we found that the design based on Akka Typed was too limiting and caused
too much boilerplate for the models we needed to implement. Essential
patterns such as request-response were found to be hard to implement
with only a single mailbox. Moveover, limiting each actor's mailbox to a
single type hindered composition and often resulted in unchecked casts
or the type being changed to `Any`, eliminating the type-safety of the
API.
In this revised API design, a simulation is now represented as the interplay of
logical processes that communicate via multiple message passing channels.
We use Kotlin coroutines to describe the behavior of the processes.
The API has been design from the start to take into account
distributed/parallel simulations by disallowing messages from arbitrary
processes, which was possible in the previous design. Instead, the
'communication graph' is known during runtime as procsses must register
themselves before being able to send/receive messages to/from channels.
We are still figuring out process/channel identity and supervision. Currently,
all logical processes run on a single level, instead of being
hierachical. However, this might change in the future.
|
|
This change renames the main module of the odcsim library to odcsim-api,
since it mainly contains the interfaces to be used by consumers of the
API and implemented by the various frameworks.
|
|
|
|
|
|
|
|
This change ensures that the batch file for invoking the Gradle wrapper
is treated as a file with Windows line endings.
|
|
|
|
|
|
This change gives an overhaul to the README for the release of version
2.x of the simulator.
|
|
This change adds the initial version of the port of the OpenDC
simulation model to version 2.x of the simulator.
The simulation model has been reworked to support immutability and
event-driven simulation, with speed-ups up to 75x.
|
|
This change adds a helper method for ActorContext to unsafely cast it to
an ActorContext of different type.
|
|
This change fixes the bug where suspending immediately after launching
the coroutine, without replacing the current behavior causes a failure
in the BehaviorInterpreter. This is because the initial behavior was not
properly set.
|
|
|
|
This change adds a StashBuffer to stash messages temporarily for
processing at a later moment.
|
|
This change makes the behavior for spawning and stopping a child actor
more consistent and better specified in the documentation.
|
|
This change adds support for spawning anonymous children in an
ActorContext. This means a name does not have to be specified when
spawning an actor.
|
|
This change adds a Behavior constructor for delaying the construction of
another Behavior instance.
|
|
This change adds support for tracking termination of actors in the
system.
|
|
This change formats the code according to the style guide of Ktlint.
|
|
This change adds support for timers in Behavior, in order to simplify
sending messages to the `self` actor.
|
|
This change adds methods to improve the composability of multiple
Behaviors, such as joining together two behaviors. Furthermore, this
change adds methods for juggling with Behavior message types.
|
|
This change adds a testkit for synchronously testking Behavior
implementations.
|
|
This change adds an actor-specific logging instance to ActorContext,
which allows consumers to log messages via the SLF4J API.
|
|
|
|
This change adds support for terminating an ActorSystem.
|
|
This change migrates the build configuration for Gradle in Groovy to
Kotlin, where possible. The scripts in the `gradle/` directory have not
been migrated to extensive use of dynamicism.
|
|
This change moves the implementation conformance test suite from the
test sources of `odcsim-core` to a new `odcsim-engine-tests` module in
order to simplify Gradle configuration.
|
|
This change updates the build dependencies Kotlin, Dokka, Ktlint and
JUnit to more recent versions.
|
|
This change makes it mandatory to send messages to an actor via the
sender's ActorContext in order for the engine to maintain consistency
in time between actors.
|
|
This change adds the initial interface design for parallel simulations,
which requires explicit synchronization between actors in order to
establish consistent virtual time between them.
|
|
This change makes a `SuspendingBehavior` instance expose its
`SuspendingActorContext` in the coroutine context, so suspending methods
inside the coroutine can access the actor context.
|
|
This change adds conversion methods from `Number` types into `Instant`
and `Duration` instances.
|
|
This change updates the build dependencies to recent released version.
These include Kotlin 1.3.21 and ktlint Gradle plugin 7.1.0.
|
|
This change makes exception handling in actor systems consistent.
Implementors should not propagate unhandled exceptions in actors to the
(one of the) threads running the engine, but instead should stop the
offending actor and continue running.
Supervision of actors should be implemented within actor behavior
instead. Since BehaviorInterpreter will propagate exceptions thrown by
an actor, consumers of that API can use it to catch unhandled exceptions in an
actor.
|
|
This change updates JUnit dependency to version 5.3.2.
|
|
This change updates the Gradle wrapper that is shipped with the
repository to version 5.1.
|
|
This change adds support for generating documentation for all modules of
the repository using Dokka.
|
|
This change adds support for suspending behavior via Kotlin's coroutine
feature officially introduced in Kotlin 1.3. The syntax is similar to
the 1.x version of the simulator to allow for easier porting to 2.x.
|
|
This change adds support for linting the code base using ktlint and the
excellent ktlint-gradle plugin.
|
|
This change adds support for deferred and special behavior in the
Behavior API. Additionally, the new Behavior DSL is implemented using
the new primitives to make the tests cleaner.
|
|
This change updates the code base to use the official Kotlin 1.3
release.
|
|
This change creates the initial conformance test suite for the API
design and tries to specify most of the requirements and assumptions
made for implementors.
|
|
This change introduces the new API design that will be introduced in the
2.x versions of the OpenDC Simulator.
This changes focuses on simplifying simulation primitives provided by
the simulator and introduces a new concept of actors based on the model
designed by the Akka Typed project.
For now, the old simulation models have been removed from the branch,
but will be ported back as this branch is being finalized.
|
|
This change will improve exception handling of processes.
At the moment, when a process throws an uncaught exception, the kernel will catch and log the exception, stop the offending process and then continue. This approach however might cause the user to overlook possible exceptions in processes and does not give any ability to the user for handling these exception.
This change modifies the kernel implementation and specification such that the `step()` method (and consequently `run()`) must propagate uncaught exceptions that occur in processes. This allows the caller to control the way exceptions are handled.
|
