From: Robert Bradshaw <robertwb@math.washington.edu>
Date: Wed, 12 Jan 2011 19:53:58 +0000 (-0800)
Subject: Fix trac #640, long string literals with escapes.
X-Git-Tag: 0.14.1rc0~23^2~2
X-Git-Url: http://git.tremily.us/?a=commitdiff_plain;h=aee29d905d4b1b753eaa927d036713d0ea9a0f16;p=cython.git

Fix trac #640, long string literals with escapes.
---

diff --git a/Cython/Compiler/StringEncoding.py b/Cython/Compiler/StringEncoding.py
index 8c474f42..fb87d4ae 100644
--- a/Cython/Compiler/StringEncoding.py
+++ b/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)