Prevent crashing from UnicodeDecodeError

nose2/plugins/buffer.py

@@ -15,6 +15,7 @@
 """
 
 import sys
+import traceback
 
 from six import StringIO
 
@@ -97,13 +98,39 @@ def outcomeDetail(self, event):
         """Add buffered output to event.extraDetail"""
         for stream in ('stdout', 'stderr'):
             if stream in event.outcomeEvent.metadata:
-                buf = event.outcomeEvent.metadata[stream].getvalue()
-                if not buf:
+                buf = ''
+                stream_buffer_exc_info = None
+                try:
+                    buf = event.outcomeEvent.metadata[stream].getvalue()
+                except UnicodeError:
+                    # python2's StringIO.StringIO [1] class has this warning:
+                    #
+                    #     The StringIO object can accept either Unicode or 8-bit strings,
+                    #     but mixing the two may take some care. If both are used, 8-bit
+                    #     strings that cannot be interpreted as 7-bit ASCII (that use the
+                    #     8th bit) will cause a UnicodeError to be raised when getvalue()
+                    #     is called.
+                    #
+                    # This exception handler is a protection against crashes
+                    # caused by this exception (such as [2] in the original
+                    # nose application). Capturing the exception info allows us
+                    # to display it back to the user.
+                    #
+                    # [1] <https://github.com/python/cpython/blob/2.7/Lib/StringIO.py#L258>
+                    # [2] <https://github.com/nose-devs/nose/issues/816>
+                    stream_buffer_exc_info = sys.exc_info()
+                if (not buf) and (not stream_buffer_exc_info):
                     continue
                 event.extraDetail.append(
                     ln('>> begin captured %s <<' % stream))
                 event.extraDetail.append(buf)
                 event.extraDetail.append(ln('>> end captured %s <<' % stream))
+                if stream_buffer_exc_info:
+                    event.extraDetail.append('OUTPUT ERROR: Could not get captured %s output.' % stream)
+                    event.extraDetail.append('The test may have output both non-ASCII Unicode strings and non-ASCII 8-bit strings.')
+                    event.extraDetail.append(ln('>> begin captured %s exception traceback <<' % stream))
+                    event.extraDetail.append(''.join(traceback.format_exception(*stream_buffer_exc_info)))
+                    event.extraDetail.append(ln('>> end captured %s exception traceback <<' % stream))
 
     def beforeInteraction(self, event):
         """Stop buffering so users can see stdout"""

nose2/tests/unit/test_buffer_plugin.py

@@ -1,9 +1,11 @@
+# -*- coding: utf-8 -*-
+
 import sys
 
 import six
 
 from nose2.plugins import buffer
-from nose2 import events, result, session
+from nose2 import events, result, session, util
 from nose2.tests._common import TestCase
 
 
@@ -18,12 +20,22 @@ def setUp(self):
 
         class Test(TestCase):
 
+            unicode_string = util.safe_decode("test 日本")
+
             def test_out(self):
                 six.print_("hello")
                 raise {}["oops"]
 
             def test_err(self):
                 six.print_("goodbye", file=sys.stderr)
+
+            def test_nonascii_mixed_unicode_and_str(self):
+                six.print_(self.unicode_string)
+                six.print_(self.unicode_string.encode('utf-8'))
+                six.print_(self.unicode_string, file=sys.stderr)
+                six.print_(self.unicode_string.encode('utf-8'), file=sys.stderr)
+                raise {}["oops"]
+
         self.case = Test
 
         class Watcher(events.Plugin):
@@ -63,6 +75,20 @@ def test_captures_stderr_when_configured(self):
         finally:
             sys.stderr = err
 
+    def test_does_not_crash_with_nonascii_mixed_unicode_and_str(self):
+        self.plugin.captureStderr = True
+        test = self.case('test_nonascii_mixed_unicode_and_str')
+        test(self.result)
+        evt = events.OutcomeDetailEvent(self.watcher.events[0])
+        self.session.hooks.outcomeDetail(evt)
+        extraDetail = "".join(evt.extraDetail)
+        if six.PY2:
+            assert self.case.unicode_string not in extraDetail, "Output unexpectedly found in error message"
+            assert "OUTPUT ERROR" in extraDetail
+            assert "UnicodeDecodeError" in extraDetail
+        else:
+            assert self.case.unicode_string in extraDetail, "Output not found in error message"
+
     def test_decorates_outcome_detail(self):
         test = self.case('test_out')
         test(self.result)