typesafehub / activator-akka-stream-scala Goto Github PK

Was implemented in Java here typesafehub/activator-akka-stream-java8@c8e514c

We need to update this template to akka-stream 1.0-m4

As seen here: #36 (comment)
with the new GroupBy, it's harder to keep it streaming all the way (i.e. submited PR went the easy way of accumulating into a List, while we must do so into a stream)

I tried to simply change the dependency version, didn't work out of the box. I'm having some headache with 2.1.0 and this is the only full example out in the wild. Documentation provides just snippets.

I haven't used Scalariform before, but it seems like meaningful assumption that sbt compile would not change the sources, from what's in the repo. Now it does.

1. git clone
2. sbt compile

$ git diff
diff --git a/src/main/scala/sample/stream/GroupLogFile.scala b/src/main/scala/sample/stream/GroupLogFile.scala
index d886c44..dcc7477 100644
--- a/src/main/scala/sample/stream/GroupLogFile.scala
+++ b/src/main/scala/sample/stream/GroupLogFile.scala
@@ -30,8 +30,8 @@ object GroupLogFile {
       via(Framing.delimiter(ByteString(System.lineSeparator), maximumFrameLength = 512, allowTruncation = true)).
       map(_.utf8String).
       map {
-        case line@LoglevelPattern(level) => (level, line)
-        case line@other => ("OTHER", line)
+        case line @ LoglevelPattern(level) => (level, line)
+        case line @ other                  => ("OTHER", line)
       }.
       // group them by log level
       groupBy(5, _._1).

It closes the file upon completion of the master stream, which has of course nothing to do with the completion of the individual groups writing (the master completes earlier). The proper solution is instead of using onComplete on the individual streams, just return the future, and call mapAsync(identity) on the master stream, and then register the onComplete(close()) call.

There's a reference in the tutorial to Transformer and Flow, but in scaladsl2, Flow does not take Transformer. In scaladsl, there's a reference to Flow.transform() but it's not clear how that connects, and there's no Transformer import.

https://github.com/typesafehub/activator-akka-stream-scala/blob/master/tutorial/index.html#L111

http://doc.akka.io/api/akka-stream-and-http-experimental/0.9/index.html#akka.stream.Transformer

=== wrong activator delete this, sorry ===

Most probably org.reactivestreams.api was renamed org.reactivestreams, this should also be done for the tutorial, e.g. on the first page in org.reactivestreams.api.Producer.

I'm going to use the Akka Streams sample in our Helsinki Scala Club meetup. Unfortunately, one of the samples reads a resource using a build-time relative path. This only works in development, of course, and not in production. In fact, the current way is not using the file as a resource, at all. All of this might be misleading to Scala learners.

Unfortunately, I don't know how to fix this for real. Using a Stream instead of File changes the later iteration code.

Here is a code that is somewhere in between. Creates a File out of a resource. This is done pretty often, but is not ideal, or recommended. Better than the current way, though.

    // Note: Resources should be read using '.getResourceAsStream'. While the code shows Akka Streams, it should not
    //      teach bad (nonfunctional, in production) manners. AKa240116
    //
    //val log: InputStream = getClass.getResourceAsStream("/logfile.txt")

    // read lines from a log file
    val logFile: File = new File(getClass.getResource("/logfile.txt").toURI)

https://github.com/typesafehub/activator-akka-stream-scala/blob/master/src/main/scala/sample/stream/WritePrimes.scala

at line 47 and 48 I have this error
value ~> is not a member of akka.stream.scaladsl.Source[Int, akka.NotUsed]

When I run WritePrimes.scala the numbers are appearing on the console faster than 1 number/second.
Also when I checked primes.txt there were always less numbers than in the console output.
Is there something I am missing?

The regex does not recognize the tags and places all lines in the OTHER bucket.

Obviously caused by the Windows eol not fully matching the line separator.