Set Jvm.spawnSubprocess input pumper thread to daemon mode.

[lihaoyi]

I noticed that sometimes when I have a bunch of zombie MillServerMain processes laying around, their jstack indicates that the only non-daemon threads are as follows:

"Thread-5" #32 prio=5 os_prio=31 cpu=193.46ms elapsed=4037.87s tid=0x00007fb64eba7000 nid=0x8303 in Object.wait()  [0x0000700002e98000]
   java.lang.Thread.State: TIMED_WAITING (on object monitor)
	at java.lang.Object.wait(java.base@11.0.7/Native Method)
	- waiting on <no object reference available>
	at java.io.PipedInputStream.read(java.base@11.0.7/PipedInputStream.java:326)
	- waiting to re-lock in wait() <0x0000000701a001c0> (a java.io.PipedInputStream)
	at java.io.PipedInputStream.read(java.base@11.0.7/PipedInputStream.java:377)
	- waiting to re-lock in wait() <0x0000000701a001c0> (a java.io.PipedInputStream)
	at java.io.InputStream.read(java.base@11.0.7/InputStream.java:205)
	at mill.main.client.InputPumper.run(InputPumper.java:25)
	at java.lang.Thread.run(java.base@11.0.7/Thread.java:834)

"Thread-9" #39 prio=5 os_prio=31 cpu=187.99ms elapsed=3987.75s tid=0x00007fb646ff9800 nid=0x15403 in Object.wait()  [0x00007000031a1000]
   java.lang.Thread.State: TIMED_WAITING (on object monitor)
	at java.lang.Object.wait(java.base@11.0.7/Native Method)
	- waiting on <no object reference available>
	at java.io.PipedInputStream.read(java.base@11.0.7/PipedInputStream.java:326)
	- waiting to re-lock in wait() <0x0000000702b3d090> (a java.io.PipedInputStream)
	at java.io.PipedInputStream.read(java.base@11.0.7/PipedInputStream.java:377)
	- waiting to re-lock in wait() <0x0000000702b3d090> (a java.io.PipedInputStream)
	at java.io.InputStream.read(java.base@11.0.7/InputStream.java:205)
	at mill.main.client.InputPumper.run(InputPumper.java:25)
	at java.lang.Thread.run(java.base@11.0.7/Thread.java:834)

These clearly indicate they are being kept alive by an InputPumper thread reading from a PipedInputStream. Grepping through the source code, this was the only place I found such an occurrence, so hopefully by marking this thread as daemon it'll allow the zombie MillServerMain processes to exit cleanly

[lefou]

LGTM!

[lihaoyi]

@nafg I think this might fix the issue you were seeing