Merge remote branch 'upstream/master'

Cython/Compiler/ModuleNode.py

@@ -6,7 +6,7 @@ import cython
 from cython import set
 cython.declare(Naming=object, Options=object, PyrexTypes=object, TypeSlots=object,
                error=object, warning=object, py_object_type=object, UtilityCode=object,
-               escape_byte_string=object, EncodedString=object)
+               EncodedString=object)
 
 import os, time
 from PyrexTypes import CPtrType
@@ -26,7 +26,7 @@ from Errors import error, warning
 from PyrexTypes import py_object_type
 from Cython.Utils import open_new_file, replace_suffix
 from Code import UtilityCode
-from StringEncoding import escape_byte_string, EncodedString
+from StringEncoding import EncodedString
 
 
 def check_c_declarations_pxd(module_node):

Cython/Compiler/Parsing.py

@@ -766,17 +766,10 @@ def p_string_literal(s, kind_override=None):
                 has_non_ASCII_literal_characters = True
         elif sy == 'ESCAPE':
             if is_raw:
-                if systr == u'\\\n':
-                    chars.append(u'\\\n')
-                elif systr == u'\\\"':
-                    chars.append(u'"')
-                elif systr == u'\\\'':
-                    chars.append(u"'")
-                else:
-                    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
+                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
             else:
                 c = systr[1]
                 if c in u"01234567":

Cython/Compiler/StringEncoding.py

@@ -232,9 +232,23 @@ def escape_byte_string(s):
                 append(c)
         return join_bytes(l).decode('ISO-8859-1')
 
-def split_string_literal(s):
+def split_string_literal(s, limit=2000):
     # MSVC can't handle long string literals.
-    if len(s) < 2047:
+    if len(s) < limit:
         return s
     else:
-        return '""'.join([s[i:i+2000] for i in range(0, len(s), 2000)]).replace(r'\""', '""\\')
+        start = 0
+        chunks = []
+        while start < len(s):
+            end = start + limit
+            if len(s) > end-4 and '\\' in s[end-4:end]:
+                end -= 4 - s[end-4:end].find('\\') # just before the backslash
+                while s[end-1] == '\\':
+                    end -= 1
+                    if end == start:
+                        # must have been a long line of backslashes
+                        end = start + limit - (limit % 2) - 4
+                        break
+            chunks.append(s[start:end])
+            start = end
+        return '""'.join(chunks)