Support all types of string literals in literals.py.

1 files changed, 44 insertions, 10 deletions

diff --git a/sphinx/pycode/pgen2/literals.py b/sphinx/pycode/pgen2/literals.py
index 0b3948a5..78667df0 100644
--- a/sphinx/pycode/pgen2/literals.py
+++ b/sphinx/pycode/pgen2/literals.py
@@ -1,6 +1,8 @@
 # Copyright 2004-2005 Elemental Security, Inc. All Rights Reserved.
 # Licensed to PSF under a Contributor Agreement.
 
+# Extended to handle raw and unicode literals by Georg Brandl.
+
 """Safely evaluate Python string literals without using eval()."""
 
 import re
@@ -16,28 +18,60 @@ simple_escapes = {"a": "\a",
                   '"': '"',
                   "\\": "\\"}
 
+def convert_hex(x, n):
+    if len(x) < n+1:
+        raise ValueError("invalid hex string escape ('\\%s')" % x)
+    try:
+        return int(x[1:], 16)
+    except ValueError:
+        raise ValueError("invalid hex string escape ('\\%s')" % x)
+
 def escape(m):
     all, tail = m.group(0, 1)
     assert all.startswith("\\")
     esc = simple_escapes.get(tail)
     if esc is not None:
         return esc
-    if tail.startswith("x"):
-        hexes = tail[1:]
-        if len(hexes) < 2:
-            raise ValueError("invalid hex string escape ('\\%s')" % tail)
+    elif tail.startswith("x"):
+        return chr(convert_hex(tail, 2))
+    elif tail.startswith('u'):
+        return unichr(convert_hex(tail, 4))
+    elif tail.startswith('U'):
+        return unichr(convert_hex(tail, 8))
+    elif tail.startswith('N'):
+        import unicodedata
         try:
-            i = int(hexes, 16)
-        except ValueError:
-            raise ValueError("invalid hex string escape ('\\%s')" % tail)
+            return unicodedata.lookup(tail[1:-1])
+        except KeyError:
+            raise ValueError("undefined character name %r" % tail[1:-1])
     else:
         try:
-            i = int(tail, 8)
+            return chr(int(tail, 8))
         except ValueError:
             raise ValueError("invalid octal string escape ('\\%s')" % tail)
-    return chr(i)
+
+def escaperaw(m):
+    all, tail = m.group(0, 1)
+    if tail.startswith('u'):
+        return unichr(convert_hex(tail, 4))
+    elif tail.startswith('U'):
+        return unichr(convert_hex(tail, 8))
+    else:
+        return all
+
+escape_re = re.compile(r"\\(\'|\"|\\|[abfnrtv]|x.{0,2}|[0-7]{1,3})")
+uni_escape_re = re.compile(r"\\(\'|\"|\\|[abfnrtv]|x.{0,2}|[0-7]{1,3}|"
+                           r"u[0-9a-fA-F]{0,4}|U[0-9a-fA-F]{0,8}|N\{.+?\})")
 
 def evalString(s):
+    regex = escape_re
+    repl = escape
+    if s.startswith('u') or s.startswith('U'):
+        regex = uni_escape_re
+        s = s[1:]
+    if s.startswith('r') or s.startswith('R'):
+        repl = escaperaw
+        s = s[1:]
     assert s.startswith("'") or s.startswith('"'), repr(s[:1])
     q = s[0]
     if s[:3] == q*3:
@@ -45,7 +79,7 @@ def evalString(s):
     assert s.endswith(q), repr(s[-len(q):])
     assert len(s) >= 2*len(q)
     s = s[len(q):-len(q)]
-    return re.sub(r"\\(\'|\"|\\|[abfnrtv]|x.{0,2}|[0-7]{1,3})", escape, s)
+    return regex.sub(repl, s)
 
 def test():
     for i in range(256):