Resolve unicode escapes in Python 2 raw unicode strings.

Closes #1594.

CHANGES.rst

@@ -27,6 +27,10 @@ Bugs fixed
 * f-string processing was adapted to match recent changes in PEP 498 and
   CPython 3.6.
 
+* Unicode escapes in 'ur' raw-unicode strings were not resolved in Py2 code.
+  Original patch by Aaron Gallagher (Github issue #1594).
+
+
 0.25.2 (2016-12-08)
 ===================
 

Cython/Compiler/Parsing.py

@@ -921,7 +921,8 @@ def p_string_literal(s, kind_override=None):
             if is_python3_source and not has_non_ascii_literal_characters and check_for_non_ascii_characters(systr):
                 has_non_ascii_literal_characters = True
         elif sy == 'ESCAPE':
-            if is_raw:
+            # in Py2, 'ur' raw unicode strings resolve unicode escapes but nothing else
+            if is_raw and (is_python3_source or kind != 'u' or systr[1] not in u'Uu'):
                 chars.append(systr)
                 if is_python3_source and not has_non_ascii_literal_characters and check_for_non_ascii_characters(systr):
                     has_non_ascii_literal_characters = True

tests/run/future_unicode_literals.pyx

@@ -9,6 +9,8 @@ if sys.version_info[0] >= 3:
     True
     >>> isinstance(b, bytes)
     True
+    >>> raw ==  'abc\\\\xf8\\\\t\\u00f8\\U000000f8'  # unescaped by Python (required by doctest)
+    True
 """
 else:
     __doc__ = u"""
@@ -18,9 +20,13 @@ else:
     True
     >>> isinstance(b, str)
     True
+    >>> raw == u'abc\\\\xf8\\\\t\\u00f8\\U000000f8'  # unescaped by Python (required by doctest)
+    True
 """
 
 u = "test"
 
 cdef char* s = "bytes test"
 b = s
+
+raw = r'abc\xf8\t\u00f8\U000000f8'

tests/run/unicodeliterals.pyx

@@ -77,6 +77,8 @@ __doc__ = br"""
     True
     >>> k == u'\\N{SNOWMAN}' == u'\\u2603'
     True
+    >>> m == u'abc\\\\xf8\\\\t\\u00f8\\U000000f8'  # unescaped by Python (required by doctest)
+    True
     >>> add == u'Søk ik' + u'üÖä' + 'abc'
     True
     >>> null == u'\\x00' # unescaped by Python (required by doctest)
@@ -110,6 +112,7 @@ f = u'\xf8'
 g = u'\udc00'   # lone trail surrogate
 h = u'\ud800'   # lone lead surrogate
 k = u'\N{SNOWMAN}'
+m = ur'abc\xf8\t\u00f8\U000000f8'
 
 add = u'Søk ik' + u'üÖä' + u'abc'
 null = u'\x00'