Better support for unicode cache entries

Author: sbidoul

src/pip/_internal/cache.py

@@ -34,7 +34,7 @@ def _hash_dict(d):
     for k in sorted(d.keys()):
         h.update(k.encode())
         h.update("=".encode())
-        h.update(d[k].encode())
+        h.update(d[k].encode("utf-8"))
         h.update(b"\0")
     return h.hexdigest()
 

tests/unit/test_cache.py

@@ -49,6 +49,8 @@ def test_cache_hash():
     assert h == "c7d60d08b1079254d236e983501fa26c016d58d16010725b27ed0af2"
     h = _hash_dict({"url": "https://g.c/o/r", "subdirectory": "sd"})
     assert h == "9cba35d4ccf04b7cde751b44db347fd0f21fa47d1276e32f9d47864c"
+    h = _hash_dict({"subdirectory": u"/\xe9e"})
+    assert h == "727d53e3305684344078965b1c0771f57f180c00516aaeb6ed8b36ce"
 
 
 def test_get_path_for_link_legacy(tmpdir):