Hi Ron, A possible recommendation is to use maven for the entire process (avoiding the sbt artifacts/processing). IJ is pretty solid in its maven support.
a) mvn -DskipTests -Pyarn -Phive -Phadoop-2.3 compile package b) Inside IJ: Open the parent/root pom.xml as a new maven project c) Inside IJ: Build | Project d) Enjoy running/debugging the individual scalatest classes e) (No unnecessary recompilation pain after the initial build) Also i found that when using maven there are no duplicate class directories (whereas in sbt everything is doubled up under the project/target directory for some reason). 2014-08-11 12:57 GMT-07:00 Ron's Yahoo! <zlgonza...@yahoo.com.invalid>: > Hi, > I’ve been able to get things compiled on my environment, but I’m > noticing that it’s been quite difficult in IntelliJ. It always recompiles > everything when I try to run one test like BroadcastTest, for example, > despite having compiled make-distribution previously. In eclipse, I have no > such recompilation issues. IntelliJ unfortunately does not support auto > compilation for Scala. It also doesn’t seem as if IntelliJ knows that that > there are classes that have already been compiled since it always opts to > recompile everything. I’m new to IntelliJ so it might really just be a lack > of knowledge on my part. > Can anyone share any tips on how they are productive compiling against > the Spark code base using IntelliJ? > > Thanks, > Ron > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@spark.apache.org > For additional commands, e-mail: dev-h...@spark.apache.org > >
