* Add test for how spaces are attached by the parser.

tests/parser/test_space_attachment.py

@@ -0,0 +1,12 @@
+from __future__ import unicode_literals
+
+import pytest
+
+@pytest.mark.models
+def test_space_attachment(EN):
+    sentence = 'This is a test.\nTo ensure  spaces are attached well.'
+    doc = EN(sentence)
+
+    for word in doc:
+        if word.is_space:
+            assert word.head.i == (word.i - 1)