* fix: keep typed ask deadletter same as classic
* fix: compile on scala3
* fix: a more convincing test
* chore: remove unnecessary header
* chore: use raw apache license
* chore: grammar and sort imports
* chore: grammar fix
* chore: copyright
* fix: immutable object domain
* fix: make it simple
- Use InteractionPatternsSpec as an example for applying Scala 3's
Union types to symplify actor code
- Remove response message wrapper & adapter
- Use the union of the actor's public protocol (the `Translate`
message (the only member of the `Command` ADT) and the possible
responses from the backend (messages `JobStarted`, `JobProgress`,
and `JobCompleted`): `private type CommandAndResponse = Command | Backend.Response`
- Instead of utilising the message adaptor `ActorRef` in the `replyTo`
field of the message sent to the backend, `context.self` is used
instead
- The internal (extended) `Behavior[CommandAndResponse]` is narrowed
to `Behavior[Command]` at creation time
The diffs between the original and the new version:
```bash
136d139
< private final case class WrappedBackendResponse(response: Backend.Response) extends Command
137a141,142
> private type CommandAndResponse = Command | Backend.Response
>
139,141c144
< Behaviors.setup[Command] { context =>
< val backendResponseMapper: ActorRef[Backend.Response] =
< context.messageAdapter(rsp => WrappedBackendResponse(rsp))
---
> Behaviors.setup[CommandAndResponse] { context =>
143,144c146,147
< def active(inProgress: Map[Int, ActorRef[URI]], count: Int): Behavior[Command] = {
< Behaviors.receiveMessage[Command] {
---
> def active(inProgress: Map[Int, ActorRef[URI]], count: Int): Behavior[CommandAndResponse] = {
> Behaviors.receiveMessage[CommandAndResponse] {
147c150
< backend ! Backend.StartTranslationJob(taskId, site, backendResponseMapper)
---
> backend ! Backend.StartTranslationJob(taskId, site, context.self)
150,162c153,162
< case wrapped: WrappedBackendResponse =>
< wrapped.response match {
< case Backend.JobStarted(taskId) =>
< context.log.info("Started {}", taskId)
< Behaviors.same
< case Backend.JobProgress(taskId, progress) =>
< context.log.info2("Progress {}: {}", taskId, progress)
< Behaviors.same
< case Backend.JobCompleted(taskId, result) =>
< context.log.info2("Completed {}: {}", taskId, result)
< inProgress(taskId) ! result
< active(inProgress - taskId, count)
< }
---
> case Backend.JobStarted(taskId) =>
> context.log.info("Started {}", taskId)
> Behaviors.same
> case Backend.JobProgress(taskId, progress) =>
> context.log.info2("Progress {}: {}", taskId, progress)
> Behaviors.same
> case Backend.JobCompleted(taskId, result) =>
> context.log.info2("Completed {}: {}", taskId, result)
> inProgress(taskId) ! result
> active(inProgress - taskId, count)
167c167
< }
---
> }.narrow
```
* Create MapAsyncPartition.scala
add license
add test
Update LICENSE
try to fix test
* Update MapAsyncPartitionSpec.scala
wip
Update MapAsyncPartition.scala
wip
* changes to get code to compile with scala 2.12
* more tests
scalafmt
* Update MapAsyncPartition.scala
* make code more closely match the Akka API
java
Update Flow.scala
more java api
* Add ordered version of the operator
* Fix formatting
* update docs
* test null function result
* java api
* add back code to get scala 2.12 compile working again
* Unify mapAsyncPartitioned implementations
* remove special license
Update CopyrightHeader.scala
* java tests
javafmt
* update docs
update tests
update javadoc
---------
Co-authored-by: Jacek Sokol <jacek@scalabs.pl>
* !remoting Update classic transport from Netty 3 to netty4
* use lambda
* =sbt Update Netty4 version to 4.1.97.final
* Reduce allocation in ChannelLocalActor.
* Remove the duplicated code in NettyHelpers.
This will make the source of errors in steps clearer, without adding any extra pipeline time.
steps in github jobs are dependent on one another, so the build files can be passed onto prValidation without the need for PR validation to compile again
