Merge pull request #1930 from rmmh/gubernator-update

Make Gubernator handle corrupt JUnit files better.
kubernetes · Feb 18, 2017 · 9f9e7af · 9f9e7af
2 parents e6565b9 + b14d264
commit 9f9e7af
Show file tree

Hide file tree

Showing 2 changed files with 25 additions and 1 deletion.
diff --git a/gubernator/view_build.py b/gubernator/view_build.py
@@ -30,7 +30,15 @@
 
 def parse_junit(xml, filename):
     """Generate failed tests as a series of (name, duration, text, filename) tuples."""
-    tree = ET.fromstring(xml)
+    try:
+        tree = ET.fromstring(xml)
+    except ET.ParseError, e:
+        logging.exception('parse_junit failed for %s', filename)
+        try:
+            tree = ET.fromstring(re.sub(r'[\x00\x80-\xFF]+', '?', xml))
+        except ET.ParseError, e:
+            yield 'Gubernator Internal Fatal XML Parse Error', 0.0, str(e), filename
+            return
     if tree.tag == 'testsuite':
         for child in tree:
             name = child.attrib['name']

diff --git a/gubernator/view_build_test.py b/gubernator/view_build_test.py
@@ -58,6 +58,22 @@ def test_testsuites(self):
     def test_bad_xml(self):
         self.assertEqual(self.parse('''<body />'''), [])
 
+    def test_corrupt_xml(self):
+        self.assertEqual(self.parse('<a>\xff</a>'), [])
+        failures = self.parse('''
+            <testsuites>
+                <testsuite name="a">
+                    <testcase name="Corrupt" time="0">
+                        <failure>something bad \xff</failure>
+                    </testcase>
+                </testsuite>
+            </testsuites>''')
+        self.assertEqual(failures, [('a Corrupt', 0.0, 'something bad ?', 'fp')])
+
+    def test_not_xml(self):
+        failures = self.parse('\x01')
+        self.assertEqual(failures,
+            [(failures[0][0], 0.0, 'not well-formed (invalid token): line 1, column 0', 'fp')])
 
 class BuildTest(main_test.TestBase):
     # pylint: disable=too-many-public-methods