Show regression vs. always-failed in cloud results

britney.py

@@ -1894,8 +1894,13 @@ class Britney(object):
                         url = cloud_url % {'h': srchash(testsrc), 's': testsrc,
                                            'r': self.options.series, 'a': arch}
                         try:
-                            r = autopkgtest.test_results[testsrc][arch][1][testver][0]
-                            status = r and 'PASS' or 'REGRESSION'
+                            if autopkgtest.test_results[testsrc][arch][1][testver][0]:
+                                status = 'PASS'
+                            else:
+                                if autopkgtest.test_results[testsrc][arch][2]:
+                                    status = 'REGRESSION'
+                                else:
+                                    status = 'ALWAYSFAIL'
                         except KeyError:
                             try:
                                 autopkgtest.pending_tests[testsrc][testver][arch]

tests/test_autopkgtest.py

@@ -256,7 +256,7 @@ lightgreen 1 i386 green 2
             # their results, we don't expect this to hold back stuff.
             VALID_CANDIDATE,
             [r'\bgreen\b.*>1</a> to .*>2<',
-             r'autopkgtest for green 2: .*amd64.*Regression.*i386.*Pass',
+             r'autopkgtest for green 2: .*amd64.*Always failed.*i386.*Pass',
              r'autopkgtest for lightgreen 1: .*amd64.*Regression.*i386.*in progress',
              r'autopkgtest for darkgreen 1: .*amd64.*in progress.*i386.*Pass'])