diff --git a/Lib/test/test_bz2.py b/Lib/test/test_bz2.py index f4e81db6e25..912fac1c334 100644 --- a/Lib/test/test_bz2.py +++ b/Lib/test/test_bz2.py @@ -47,6 +47,7 @@ class BaseTest(unittest.TestCase): ] TEXT = b''.join(TEXT_LINES) DATA = b'BZh91AY&SY.\xc8N\x18\x00\x01>_\x80\x00\x10@\x02\xff\xf0\x01\x07n\x00?\xe7\xff\xe00\x01\x99\xaa\x00\xc0\x03F\x86\x8c#&\x83F\x9a\x03\x06\xa6\xd0\xa6\x93M\x0fQ\xa7\xa8\x06\x804hh\x12$\x11\xa4i4\xf14S\xd2\x88\xe5\xcd9gd6\x0b\n\xe9\x9b\xd5\x8a\x99\xf7\x08.K\x8ev\xfb\xf7xw\xbb\xdf\xa1\x92\xf1\xdd|/";\xa2\xba\x9f\xd5\xb1#A\xb6\xf6\xb3o\xc9\xc5y\\\xebO\xe7\x85\x9a\xbc\xb6f8\x952\xd5\xd7"%\x89>V,\xf7\xa6z\xe2\x9f\xa3\xdf\x11\x11"\xd6E)I\xa9\x13^\xca\xf3r\xd0\x03U\x922\xf26\xec\xb6\xed\x8b\xc3U\x13\x9d\xc5\x170\xa4\xfa^\x92\xacDF\x8a\x97\xd6\x19\xfe\xdd\xb8\xbd\x1a\x9a\x19\xa3\x80ankR\x8b\xe5\xd83]\xa9\xc6\x08\x82f\xf6\xb9"6l$\xb8j@\xc0\x8a\xb0l1..\xbak\x83ls\x15\xbc\xf4\xc1\x13\xbe\xf8E\xb8\x9d\r\xa8\x9dk\x84\xd3n\xfa\xacQ\x07\xb1%y\xaav\xb4\x08\xe0z\x1b\x16\xf5\x04\xe9\xcc\xb9\x08z\x1en7.G\xfc]\xc9\x14\xe1B@\xbb!8`' + EMPTY_DATA = b'BZh9\x17rE8P\x90\x00\x00\x00\x00' def setUp(self): self.filename = TESTFN @@ -584,6 +585,12 @@ class BZ2CompressorTest(BaseTest): data += bz2c.flush() self.assertEqual(self.decompress(data), self.TEXT) + def testCompressEmptyString(self): + bz2c = BZ2Compressor() + data = bz2c.compress(b'') + data += bz2c.flush() + self.assertEqual(data, self.EMPTY_DATA) + def testCompressChunks10(self): bz2c = BZ2Compressor() n = 0 @@ -671,6 +678,10 @@ class CompressDecompressTest(BaseTest): data = bz2.compress(self.TEXT) self.assertEqual(self.decompress(data), self.TEXT) + def testCompressEmptyString(self): + text = bz2.compress(b'') + self.assertEqual(text, self.EMPTY_DATA) + def testDecompress(self): text = bz2.decompress(self.DATA) self.assertEqual(text, self.TEXT) @@ -679,6 +690,10 @@ class CompressDecompressTest(BaseTest): text = bz2.decompress(b"") self.assertEqual(text, b"") + def testDecompressToEmptyString(self): + text = bz2.decompress(self.EMPTY_DATA) + self.assertEqual(text, b'') + def testDecompressIncomplete(self): self.assertRaises(ValueError, bz2.decompress, self.DATA[:-10]) diff --git a/Misc/ACKS b/Misc/ACKS index 36aed17fe3d..8dbdc4dc208 100644 --- a/Misc/ACKS +++ b/Misc/ACKS @@ -884,6 +884,7 @@ Michael Otteneder R. M. Oudkerk Russel Owen Joonas Paalasmaa +Martin Packman Shriphani Palakodety Ondrej Palkovsky Mike Pall diff --git a/Misc/NEWS b/Misc/NEWS index fdc313914e3..eff20ee93e1 100644 --- a/Misc/NEWS +++ b/Misc/NEWS @@ -123,6 +123,9 @@ Core and Builtins Library ------- +- Issue #16828: Fix error incorrectly raised by bz2.compress(''). Initial patch + by Martin Packman. + - Issue #16541: tk_setPalette() now works with keyword arguments. - Issue #16820: In configparser, `parser.popitem()` no longer raises ValueError. diff --git a/Modules/_bz2module.c b/Modules/_bz2module.c index 5cac8e6d518..4eee5a2feeb 100644 --- a/Modules/_bz2module.c +++ b/Modules/_bz2module.c @@ -145,34 +145,24 @@ compress(BZ2Compressor *c, char *data, size_t len, int action) result = PyBytes_FromStringAndSize(NULL, SMALLCHUNK); if (result == NULL) return NULL; + c->bzs.next_in = data; - /* On a 64-bit system, len might not fit in avail_in (an unsigned int). - Do compression in chunks of no more than UINT_MAX bytes each. */ - c->bzs.avail_in = MIN(len, UINT_MAX); - len -= c->bzs.avail_in; + c->bzs.avail_in = 0; c->bzs.next_out = PyBytes_AS_STRING(result); c->bzs.avail_out = PyBytes_GET_SIZE(result); for (;;) { char *this_out; int bzerror; - Py_BEGIN_ALLOW_THREADS - this_out = c->bzs.next_out; - bzerror = BZ2_bzCompress(&c->bzs, action); - data_size += c->bzs.next_out - this_out; - Py_END_ALLOW_THREADS - if (catch_bz2_error(bzerror)) - goto error; - + /* On a 64-bit system, len might not fit in avail_in (an unsigned int). + Do compression in chunks of no more than UINT_MAX bytes each. */ if (c->bzs.avail_in == 0 && len > 0) { c->bzs.avail_in = MIN(len, UINT_MAX); len -= c->bzs.avail_in; } - /* In regular compression mode, stop when input data is exhausted. - In flushing mode, stop when all buffered data has been flushed. */ - if ((action == BZ_RUN && c->bzs.avail_in == 0) || - (action == BZ_FINISH && bzerror == BZ_STREAM_END)) + /* In regular compression mode, stop when input data is exhausted. */ + if (action == BZ_RUN && c->bzs.avail_in == 0) break; if (c->bzs.avail_out == 0) { @@ -185,6 +175,18 @@ compress(BZ2Compressor *c, char *data, size_t len, int action) } c->bzs.avail_out = MIN(buffer_left, UINT_MAX); } + + Py_BEGIN_ALLOW_THREADS + this_out = c->bzs.next_out; + bzerror = BZ2_bzCompress(&c->bzs, action); + data_size += c->bzs.next_out - this_out; + Py_END_ALLOW_THREADS + if (catch_bz2_error(bzerror)) + goto error; + + /* In flushing mode, stop when all buffered data has been flushed. */ + if (action == BZ_FINISH && bzerror == BZ_STREAM_END) + break; } if (data_size != PyBytes_GET_SIZE(result)) if (_PyBytes_Resize(&result, data_size) < 0)