Merged in jaapz/pytest-xml-escape-control-chars (pull request #32)

Fix junitxml generation when using special characters in parametrized tests.

_pytest/junitxml.py

@@ -36,7 +36,8 @@ class Junit(py.xml.Namespace):
 #                    | [#x10000-#x10FFFF]
 _legal_chars = (0x09, 0x0A, 0x0d)
 _legal_ranges = (
-    (0x20, 0xD7FF),
+    (0x20, 0x7E),
+    (0x80, 0xD7FF),
     (0xE000, 0xFFFD),
     (0x10000, 0x10FFFF),
 )
@@ -103,7 +104,7 @@ class LogXML(object):
             classnames.insert(0, self.prefix)
         self.tests.append(Junit.testcase(
             classname=".".join(classnames),
-            name=names[-1],
+            name=bin_xml_escape(names[-1]),
             time=getattr(report, 'duration', 0)
         ))
 

testing/test_junitxml.py

@@ -450,3 +450,16 @@ def test_logxml_changingdir(testdir):
     assert result.ret == 0
     assert testdir.tmpdir.join("a/x.xml").check()
 
+def test_escaped_parametrized_names_xml(testdir):
+    testdir.makepyfile("""
+        import pytest
+        @pytest.mark.parametrize('char', ["\\x00"])
+        def test_func(char):
+            assert char
+    """)
+    result, dom = runandparse(testdir)
+    assert result.ret == 0
+    node = dom.getElementsByTagName("testcase")[0]
+    assert_attr(node,
+        name="test_func[#x00]")
+