Timing window in sys.process.ProcessImpl.scala

I've noticed a failure running some Apache Spark CLI unit tests.

In class PipedProcesses we indicate the input stream has been read into a 
PipeInputStream on line 138 by "currentSink put None".

However if this occurs before the command has got going ( referred to in 
the code as second ) it can never see the input stream. The run method in 
class PipeSink can just fall through to the case None ( line 183 ) without 
any input having been read.

I believe it can be simply fixed by adding and checking a new SyncVar ( 
maybe called readComplete ) which is set once the runloop has completed 
reading data from the pipe.

Is it appropriate if I just create a Scala issue? Sorry if that's a dumb 
question - new to the Scala community.

-- 
You received this message because you are subscribed to the Google Groups "scala-language" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.