From: Hye-Shik Chang <hyeshik@gmail.com>
Date: Tue, 5 Jun 2007 19:14:33 +0000 (+0000)
Subject: (Backport from r55770)
X-Git-Tag: v2.5.2c1~279
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=60111e09a915be77762c7af5448ea34a61de4577;p=python

(Backport from r55770)

Bug #1728403: Fix a bug that CJKCodecs StreamReader hangs when it
reads a file that ends with incomplete sequence and sizehint argument
for .read() is specified.
---

diff --git a/Lib/test/test_multibytecodec.py b/Lib/test/test_multibytecodec.py
index a8666d31cc..dd933f3463 100644
--- a/Lib/test/test_multibytecodec.py
+++ b/Lib/test/test_multibytecodec.py
@@ -136,6 +136,14 @@ class Test_IncrementalDecoder(unittest.TestCase):
         self.assertRaises(UnicodeDecodeError, decoder.decode, '', True)
         self.assertEqual(decoder.decode('B@$'), u'\u4e16')
 
+class Test_StreamReader(unittest.TestCase):
+    def test_bug1728403(self):
+        try:
+            open(TESTFN, 'w').write('\xa1')
+            f = codecs.open(TESTFN, encoding='cp949')
+            self.assertRaises(UnicodeDecodeError, f.read, 2)
+        finally:
+            os.unlink(TESTFN)
 
 class Test_StreamWriter(unittest.TestCase):
     if len(u'\U00012345') == 2: # UCS2
@@ -223,6 +231,7 @@ def test_main():
     suite.addTest(unittest.makeSuite(Test_MultibyteCodec))
     suite.addTest(unittest.makeSuite(Test_IncrementalEncoder))
     suite.addTest(unittest.makeSuite(Test_IncrementalDecoder))
+    suite.addTest(unittest.makeSuite(Test_StreamReader))
     suite.addTest(unittest.makeSuite(Test_StreamWriter))
     suite.addTest(unittest.makeSuite(Test_ISO2022))
     test_support.run_suite(suite)
diff --git a/Misc/NEWS b/Misc/NEWS
index 10bf25c93d..f4be238c60 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -12,6 +12,10 @@ What's New in Python 2.5.2c1?
 Library
 -------
 
+- Bug #1728403: Fix a bug that CJKCodecs StreamReader hangs when it
+  reads a file that ends with incomplete sequence and sizehint argument
+  for .read() is specified.
+
 - Bug #1730389: Have time.strptime() match spaces in a format argument with
   ``\s+`` instead of ``\s*``.
 
diff --git a/Modules/cjkcodecs/multibytecodec.c b/Modules/cjkcodecs/multibytecodec.c
index 7c6b989355..9fb9570225 100644
--- a/Modules/cjkcodecs/multibytecodec.c
+++ b/Modules/cjkcodecs/multibytecodec.c
@@ -1214,6 +1214,8 @@ mbstreamreader_iread(MultibyteStreamReaderObject *self,
 	cres = NULL;
 
 	for (;;) {
+		int endoffile;
+
 		if (sizehint < 0)
 			cres = PyObject_CallMethod(self->stream,
 					(char *)method, NULL);
@@ -1230,6 +1232,8 @@ mbstreamreader_iread(MultibyteStreamReaderObject *self,
 			goto errorexit;
 		}
 
+		endoffile = (PyString_GET_SIZE(cres) == 0);
+
 		if (self->pendingsize > 0) {
 			PyObject *ctr;
 			char *ctrdata;
@@ -1257,7 +1261,7 @@ mbstreamreader_iread(MultibyteStreamReaderObject *self,
 				(MultibyteStatefulDecoderContext *)self, &buf))
 			goto errorexit;
 
-		if (rsize == 0 || sizehint < 0) { /* end of file */
+		if (endoffile || sizehint < 0) {
 			if (buf.inbuf < buf.inbuf_end &&
 			    multibytecodec_decerror(self->codec, &self->state,
 					&buf, self->errors, MBERR_TOOFEW))