Fixed a recompilation issue in CI, tweaked coverage.py a bit more

This was lost in the Travis -> GitHub transition, in serializing some of
the jobs, I missed that we need to clean between tests with different
geometry configurations. Otherwise we end up running outdated binaries,
which explains some of the weird test behavior we were seeing.

Also tweaked a few script things:
- Better subprocess error reporting (dump stderr on failure)
- Fixed a BUILDDIR rule issue in test.py
- Changed test-not-run status to None instead of undefined

scripts/code.py

@@ -28,11 +28,20 @@ def collect(paths, **args):
         cmd = args['nm_tool'] + ['--size-sort', path]
         if args.get('verbose'):
             print(' '.join(shlex.quote(c) for c in cmd))
-        proc = sp.Popen(cmd, stdout=sp.PIPE, universal_newlines=True)
+        proc = sp.Popen(cmd,
+            stdout=sp.PIPE,
+            stderr=sp.PIPE if not args.get('verbose') else None,
+            universal_newlines=True)
         for line in proc.stdout:
             m = pattern.match(line)
             if m:
                 results[(path, m.group('func'))] += int(m.group('size'), 16)
+        proc.wait()
+        if proc.returncode != 0:
+            if not args.get('verbose'):
+                for line in proc.stderr:
+                    sys.stdout.write(line)
+            sys.exit(-1)
 
     flat_results = []
     for (file, func), size in results.items():