Use is-reachable to detect child process health in self-test.

On Linux, child processes that have exited may remain as <defunct>
"zombie" processes, which prevents process.kill(childPid, 0) from
throwing, so we need a different trick for detecting whether the child
process is still alive.

scripts/dev-bundle-tool-package.js

@@ -41,6 +41,7 @@ var packageJson = {
     sqlite3: "3.1.8",
     netroute: "1.0.2",
     "http-proxy": "1.16.2",
+    "is-reachable": "3.1.0",
     "wordwrap": "0.0.2",
     "moment": "2.20.1",
     "rimraf": "2.6.2",

tools/tests/run.js

@@ -7,6 +7,7 @@ var _ = require('underscore');
 var files = require('../fs/files.js');
 var catalog = require('../packaging/catalog/catalog.js');
 var os = require('os');
+var isReachable = require("is-reachable");
 
 var DEFAULT_RELEASE_TRACK = catalog.DEFAULT_TRACK;
 
@@ -377,6 +378,10 @@ selftest.define("run and SIGKILL parent process", ["yet-unsolved-windows-failure
   }
   childPid = match[1];
 
+  if (!isReachable("localhost:3000").await()) {
+    selftest.fail("Child process " + childPid + " already dead?");
+  }
+
   process.kill(run.proc.pid, "SIGKILL");
   // This sleep should be a little more time than the interval at which
   // the child checks if the parent is still alive, in
@@ -386,14 +391,7 @@ selftest.define("run and SIGKILL parent process", ["yet-unsolved-windows-failure
   // Send the child process a signal of 0. If there is no error, it
   // means that the process is still running, which is not what we
   // expect.
-  var caughtError;
-  try {
-    process.kill(childPid, 0);
-  } catch (err) {
-    caughtError = err;
-  }
-
-  if (! caughtError) {
+  if (isReachable("localhost:3000").await()) {
     selftest.fail("Child process " + childPid + " is still running");
   }