I'd like to create a dummy script (with the prospect of writing a real life script) that invokes for example the Python interactive interpreter from within a Scala process and lets the user fully interact with the subprocess; i.e. the stdin/stdout/stderr of the child process should be connected to those of the parent (Scala) process. I've tried using the following to no avail:
#!/usr/bin/env scala -savecompiled
import sys.process._
stringToProcess("python").run(BasicIO.standard(connectInput = true)).exitValue
however, while it does seem to successfully run a python subprocess, the only interaction I get to have with it is Ctrl-C:
~$ ./scalashelltest.scala
foo
^CTraceback (most recent call last):
File "<stdin>", line 1, in <module>
KeyboardInterrupt
If press Ctrl-C immediately, I don't even get that output:
~$ ./scalashelltest.scala
^C~$
Any idea why this is happening and how to make it work as expected?
You aren't giving python a (pseudo-)tty. You are just giving it stdin (and possibly stdout).
So python is operating in a non-interactive mode.
Running python with the -i flag will force it to use prompts even without a tty but the more correct fix is likely to find a scala/java library which can create a (pseudo-)tty and run an application in it.