When a test fails, run it first the next time

This patch makes run_job() return a very large "elapsed time" for jobs
that fail. The effect is that on the next run, tests that previously
failed will be tried first regardless of their actual runtime; since
tests that failed before are likely to fail again, and since observing
failures is often the reason one runs tests in the first place, this
will usually greatly reduce the time one has to wait.

gtest-parallel

diff --git a/gtest-parallel b/gtest-parallel
index 21c6ee0..8a9969e 100755
--- a/gtest-parallel
+++ b/gtest-parallel
@@ -221,7 +221,10 @@
 
 exit_code = 0
 
-# Run the specified job. Returns the elapsed time in milliseconds.
+# Run the specified job. Return the elapsed time in milliseconds if
+# the job succeeds, or a very large number (larger than any reasonable
+# elapsed time) if the job fails. (This ensures that failing tests
+# will run first the next time.)
 def run_job((command, job_id, test)):
   begin = time.time()
   sub = subprocess.Popen(command + ['--gtest_filter=' + test] +
@@ -238,10 +241,11 @@
   code = sub.wait()
   runtime_ms = int(1000 * (time.time() - begin))
   logger.log("%s: EXIT %s %d" % (job_id, code, runtime_ms))
-  if code != 0:
-    global exit_code
-    exit_code = code
-  return runtime_ms
+  if code == 0:
+    return runtime_ms
+  global exit_code
+  exit_code = code
+  return sys.maxint
 
 def worker():
   global job_id