Prevent UnicodeErrors in HTML reports with Unicode configuration values

3 files changed, 13 insertions, 3 deletions

diff --git a/CHANGES.rst b/CHANGES.rst
index 91218bb..d56aebe 100644
--- a/CHANGES.rst
+++ b/CHANGES.rst
@@ -53,6 +53,9 @@ Unreleased
 
 - Deal properly with non-ASCII file names in an ASCII-only world, `issue 533`_.
 
+- Plugins that set Unicode configuration values could cause UnicodeErrors when
+  generating HTML reports.  This is now fixed.
+
 - Prevented deprecation warnings from configparser that happened in some
   circumstances, closing `issue 530`_.
 
diff --git a/coverage/misc.py b/coverage/misc.py
index 9b1894f..5d330c6 100644
--- a/coverage/misc.py
+++ b/coverage/misc.py
@@ -12,7 +12,7 @@ import sys
 import types
 
 from coverage import env
-from coverage.backward import string_class, to_bytes, unicode_class
+from coverage.backward import to_bytes, unicode_class
 
 ISOLATED_MODULES = {}
 
@@ -179,8 +179,8 @@ class Hasher(object):
     def update(self, v):
         """Add `v` to the hash, recursively if needed."""
         self.md5.update(to_bytes(str(type(v))))
-        if isinstance(v, string_class):
-            self.md5.update(to_bytes(v))
+        if isinstance(v, unicode_class):
+            self.md5.update(v.encode('utf8'))
         elif isinstance(v, bytes):
             self.md5.update(v)
         elif v is None:
diff --git a/tests/test_misc.py b/tests/test_misc.py
index 38be345..96b5100 100644
--- a/tests/test_misc.py
+++ b/tests/test_misc.py
@@ -34,6 +34,13 @@ class HasherTest(CoverageTest):
         h2.update(b"Goodbye!")
         self.assertNotEqual(h1.hexdigest(), h2.hexdigest())
 
+    def test_unicode_hashing(self):
+        h1 = Hasher()
+        h1.update(u"Hello, world! \N{SNOWMAN}")
+        h2 = Hasher()
+        h2.update(u"Goodbye!")
+        self.assertNotEqual(h1.hexdigest(), h2.hexdigest())
+
     def test_dict_hashing(self):
         h1 = Hasher()
         h1.update({'a': 17, 'b': 23})