[FIXED HUDSON-8244] turns out Process.getErrorStream() needs to be explicitly...

[FIXED HUDSON-8244] turns out Process.getErrorStream() needs to be explicitly closed even when stderr is redirected to stdout.

changelog.html

@@ -43,6 +43,9 @@ Upcoming changes</a>
   <li class=bug>
     " (from WhateverTest)" gratuitously appended to test result detail pages.
     (<a href="http://issues.hudson-ci.org/browse/HUDSON-5655">issue 5655</a>)
+  <li class=bug>
+    Fixed a pipe leak to child processes.
+    (<a href="http://issues.hudson-ci.org/browse/HUDSON-8244">issue 8244</a>)
   <li class=rfe>
     Debian package init script now honors <tt>~/.profile</tt>.
   <li class=rfe>

core/src/main/java/hudson/Proc.java

@@ -200,6 +200,10 @@ public abstract class Proc {
                 copier2 = new StreamCopyThread(name+": stderr copier", proc.getErrorStream(), err);
                 copier2.start();
             } else {
+                // while this is not discussed in javadoc, even with ProcessBuilder.redirectErrorStream(true),
+                // Process.getErrorStream() still returns a distinct reader end of a pipe that needs to be closed.
+                // this is according to the source code of JVM
+                proc.getErrorStream().close();
                 copier2 = null;
             }
         }