Make fuzz-natives test generator more robust.

test/fuzz-natives/testcfg.py

Index: test/fuzz-natives/testcfg.py
diff --git a/test/fuzz-natives/testcfg.py b/test/fuzz-natives/testcfg.py
index df6bc5bf8776a575b153cef5d288641218525df0..5e00b404bc926a723eb44cf5b8d8bd39aed2ad8d 100644
--- a/test/fuzz-natives/testcfg.py
+++ b/test/fuzz-natives/testcfg.py
@@ -31,11 +31,16 @@ class FuzzNativesTestSuite(testsuite.TestSuite):
       assert False, "Failed to get natives list."
     tests = []
     for line in output.stdout.strip().split():
-      (name, argc) = line.split(",")
-      flags = ["--allow-natives-syntax",
-               "-e", "var NAME = '%s', ARGC = %s;" % (name, argc)]
-      test = testcase.TestCase(self, name, flags)
-      tests.append(test)
+      try:
+        (name, argc) = line.split(",")
+        flags = ["--allow-natives-syntax",
+                 "-e", "var NAME = '%s', ARGC = %s;" % (name, argc)]
+        test = testcase.TestCase(self, name, flags)
+        tests.append(test)
+      except:
+        # Work-around: If parsing didn't work, it might have been due to output
+        # caused by other d8 flags.
+        pass
     return tests
 
   def GetFlagsForTestCase(self, testcase, context):