Issue #8260: The read(), readline() and readlines() methods of
codecs.StreamReader returned incomplete data when were called after
readline() or read(size). Based on patch by Amaury Forgeot d'Arc.
diff --git a/Lib/codecs.py b/Lib/codecs.py
index f4cd60a..79a918d 100644
--- a/Lib/codecs.py
+++ b/Lib/codecs.py
@@ -456,16 +456,13 @@
# read until we get the required number of characters (if available)
while True:
- # can the request can be satisfied from the character buffer?
- if chars < 0:
- if size < 0:
- if self.charbuffer:
- break
- elif len(self.charbuffer) >= size:
- break
- else:
+ # can the request be satisfied from the character buffer?
+ if chars >= 0:
if len(self.charbuffer) >= chars:
break
+ elif size >= 0:
+ if len(self.charbuffer) >= size:
+ break
# we need more data
if size < 0:
newdata = self.stream.read()
@@ -473,6 +470,8 @@
newdata = self.stream.read(size)
# decode bytes (those remaining from the last call included)
data = self.bytebuffer + newdata
+ if not data:
+ break
try:
newchars, decodedbytes = self.decode(data, self.errors)
except UnicodeDecodeError, exc:
diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
index c9a2515..d0f96af 100644
--- a/Lib/test/test_codecs.py
+++ b/Lib/test/test_codecs.py
@@ -124,6 +124,40 @@
size*u"a",
)
+ def test_mixed_readline_and_read(self):
+ lines = ["Humpty Dumpty sat on a wall,\n",
+ "Humpty Dumpty had a great fall.\r\n",
+ "All the king's horses and all the king's men\r",
+ "Couldn't put Humpty together again."]
+ data = ''.join(lines)
+ def getreader():
+ stream = StringIO.StringIO(data.encode(self.encoding))
+ return codecs.getreader(self.encoding)(stream)
+
+ # Issue #8260: Test readline() followed by read()
+ f = getreader()
+ self.assertEqual(f.readline(), lines[0])
+ self.assertEqual(f.read(), ''.join(lines[1:]))
+ self.assertEqual(f.read(), '')
+
+ # Issue #16636: Test readline() followed by readlines()
+ f = getreader()
+ self.assertEqual(f.readline(), lines[0])
+ self.assertEqual(f.readlines(), lines[1:])
+ self.assertEqual(f.read(), '')
+
+ # Test read() followed by read()
+ f = getreader()
+ self.assertEqual(f.read(size=40, chars=5), data[:5])
+ self.assertEqual(f.read(), data[5:])
+ self.assertEqual(f.read(), '')
+
+ # Issue #12446: Test read() followed by readlines()
+ f = getreader()
+ self.assertEqual(f.read(size=40, chars=5), data[:5])
+ self.assertEqual(f.readlines(), [lines[0][5:]] + lines[1:])
+ self.assertEqual(f.read(), '')
+
def test_bug1175396(self):
s = [
'<%!--===================================================\r\n',
diff --git a/Misc/NEWS b/Misc/NEWS
index 3a2d2da..d1f7b7b 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -38,6 +38,10 @@
Library
-------
+- Issue #8260: The read(), readline() and readlines() methods of
+ codecs.StreamReader returned incomplete data when were called after
+ readline() or read(size). Based on patch by Amaury Forgeot d'Arc.
+
- Issue #20374: Fix build with GNU readline >= 6.3.
- Issue #14548: Make multiprocessing finalizers check pid before