Added script to run tests and compare their outputs on linux and fuchsia NUCs.

comparative_tester.py contains the initial comparative testing script, designed
to accept a list of test targets, compile and build the targets(and generate
them with default args using gn if necessary), build them using ninja and goma,
and then execute them, sending targets one at a time to both linux and fuchsia
NUCs to be executed. Once a test is completed, its output is filtered for
informational lines, they are collated, and written to csv.

Bug: 839491

Change-Id: I287470ae9f50f169fbd1d9296c69ee9560318837
Reviewed-on: https://chromium-review.googlesource.com/1031294
Commit-Queue: Stephan Stross <stephanstross@google.com>
Reviewed-by: Sergey Ulanov <sergeyu@chromium.org>
Cr-Commit-Position: refs/heads/master@{#558362}

tools/fuchsia/comparative_tester/target_spec.py

+# Copyright 2018 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+# Fields for use when working with a physical linux device connected locally
+linux_device_hostname = "192.168.42.32"
+linux_device_user = "potat"
+
+linux_out_dir = "out/default"
+fuchsia_out_dir = "out/fuchsia"
+
+# A map of test target names to a list of test filters to be passed to
+# --gtest_filter. Stick to *_perftests. Also, whoo implicit string
+# joining!
+test_targets = {
+    "base:base_perftests": "-WaitableEventPerfTest.Throughput"
+        ":MessageLoopPerfTest.PostTaskRate/1_Posting_Thread",
+    "net:net_perftests": "",
+}