diff --git a/Lib/gzip.py b/Lib/gzip.py index 4ecd2115ffe..3c1ebf25142 100644 --- a/Lib/gzip.py +++ b/Lib/gzip.py @@ -55,6 +55,7 @@ class GzipFile: """ myfileobj = None + max_read_chunk = 10 * 1024 * 1024 # 10Mb def __init__(self, filename=None, mode=None, compresslevel=9, fileobj=None): @@ -215,14 +216,14 @@ class GzipFile: try: while True: self._read(readsize) - readsize = readsize * 2 + readsize = min(self.max_read_chunk, readsize * 2) except EOFError: size = self.extrasize else: # just get some more of it try: while size > self.extrasize: self._read(readsize) - readsize = readsize * 2 + readsize = min(self.max_read_chunk, readsize * 2) except EOFError: if size > self.extrasize: size = self.extrasize diff --git a/Lib/test/test_gzip.py b/Lib/test/test_gzip.py index 9e5cb0370a7..0f8e03e0a45 100644 --- a/Lib/test/test_gzip.py +++ b/Lib/test/test_gzip.py @@ -58,6 +58,29 @@ class TestGzip(unittest.TestCase): f = gzip.GzipFile(self.filename, 'rb') ; d = f.read() ; f.close() self.assertEqual(d, (data1*50) + (data2*15)) + def test_many_append(self): + # Bug #1074261 was triggered when reading a file that contained + # many, many members. Create such a file and verify that reading it + # works. + f = gzip.open(self.filename, 'wb', 9) + f.write('a') + f.close() + for i in range(0,200): + f = gzip.open(self.filename, "ab", 9) # append + f.write('a') + f.close() + + # Try reading the file + zgfile = gzip.open(self.filename, "rb") + contents = "" + while 1: + ztxt = zgfile.read(8192) + contents += ztxt + if not ztxt: break + zgfile.close() + self.assertEquals(contents, 'a'*201) + + def test_readline(self): self.test_write() # Try .readline() with varying line lengths