Set an upper limit on the size of the field buffer, raise an exception

when this limit is reached. Limit defaults to 128k, and is changed
by module set_field_limit() method. Previously, an unmatched quote
character could result in the entire file being read into the field
buffer, potentially exhausting virtual memory.

2 files changed, 10 insertions, 2 deletions

diff --git a/Lib/csv.py b/Lib/csv.py
index 37bdba5..aaaa3cf 100644
--- a/Lib/csv.py
+++ b/Lib/csv.py
@@ -6,6 +6,7 @@ csv.py - read/write/investigate CSV files
 import re
 from _csv import Error, __version__, writer, reader, register_dialect, \
                  unregister_dialect, get_dialect, list_dialects, \
+                 set_field_limit, \
                  QUOTE_MINIMAL, QUOTE_ALL, QUOTE_NONNUMERIC, QUOTE_NONE, \
                  __doc__
 from _csv import Dialect as _Dialect
diff --git a/Lib/test/test_csv.py b/Lib/test/test_csv.py
index ff45b61..82a36e9 100644
--- a/Lib/test/test_csv.py
+++ b/Lib/test/test_csv.py
@@ -229,10 +229,17 @@ class Test_Csv(unittest.TestCase):
                         quoting=csv.QUOTE_NONE, escapechar='\\')
 
     def test_read_bigfield(self):
-        # This exercises the buffer realloc functionality
-        bigstring = 'X' * 50000
+        # This exercises the buffer realloc functionality and field size
+        # limits.
+        size = 50000
+        bigstring = 'X' * size
         bigline = '%s,%s' % (bigstring, bigstring)
         self._read_test([bigline], [[bigstring, bigstring]])
+        csv.set_field_limit(size)
+        self._read_test([bigline], [[bigstring, bigstring]])
+        self.assertEqual(csv.set_field_limit(), size)
+        csv.set_field_limit(size-1)
+        self.assertRaises(csv.Error, self._read_test, [bigline], [])
 
 class TestDialectRegistry(unittest.TestCase):
     def test_registry_badargs(self):