run and runAll directly stops after services where started

Lagom Version (1.2.x / 1.3.x / etc)

1.4.0

API (Scala / Java / Neither / Both)

Scala

Operating System (Ubuntu 15.10 / MacOS 10.10 / Windows 10)

macOS Darwin deess1nb1337.de.root.net 17.4.0 Darwin Kernel Version 17.4.0: Sun Dec 17 09:19:54 PST 2017; root:xnu-4570.41.2~1/RELEASE_X86_64 x86_64

JDK (Oracle 1.8.0_112, OpenJDK 1.8.x, Azul Zing)

Oracle java version “1.8.0_141”

The issue

Unfortunately it’s hard to reproduce but maybe others have the same issue. After starting/stopping Lagom a few times I can not start it anymore because it directly stops. This happens for runAll but also if I directly start infra services and start a single service with run. The output looks similar to this:

09:56:33.730 | ForkJoinPoo | INFO  | play.api.Play - Application started (Dev)
[info] Service a1-impl listening for HTTP on 0:0:0:0:0:0:0:0:62636
[info] Service a2-impl listening for HTTP on 0:0:0:0:0:0:0:0:51040
[info] Service a3-impl listening for HTTP on 0:0:0:0:0:0:0:0:49889
[info] (Services started, press enter to stop and go back to the console...)
[info] Stopping services
10:04:55.687 | pool-39-threa | INFO  | play.core.server.AkkaHttpServer - Stopping server...
10:04:55.689 [info] play.core.server.AkkaHttpServer [] - Stopping server...
10:04:55.692 | pool-39-threa | INFO  | play.core.server.AkkaHttpServer - Stopping server...
10:04:55.692 | pool-39-thread | INFO  | play.core.server.AkkaHttpServer - Stopping server...

So I do not press any key, it just directly stops. A workaround usually is to clean the whole project.

It is really annoying because it hinders productivity a lot.

[fkoehler]

Believe it or not but I was able to produce really simple basic testcase: https://github.com/fkoehler/lagom-sbt-java-console-read-bug

In that simple sbt project I just run either System.in.read() or System.console().readLine(). With System.console().readLine() everything works as expected.

With System.in.read() the call to run immediately returns the second time when I pressed ENTER already in the first run. So the output looks like this:

[info] Set current project to consoletest (in build file:/Users/fab/dev/consoletest/)
> run
[info] Running Main 
reading ....
sdfsafsdfsdf
done!
[success] Total time: 3 s, completed 27.02.2018 22:33:52
> run
[info] Running Main 
reading ....
done!
[success] Total time: 0 s, completed 27.02.2018 22:33:54

That’s exactly what I experience in Lagom. So System.console.readLine() seems to be more stable. Hard do write an automatic test I guess.

Is that enough to create a fix for this or is one person having the issue not good enough? Since it works exactly the same I think changing that one line would not be an issue? I would not mind creating a PR for that.

[fkoehler]

I can’t find anything related in terminal regarding those settings. Maybe I somehow enter that mode. I will keep investigating in that direction.