Skip to content

Commit

Permalink
pythongh-101961 fileinput.hookcompressed should not set the encoding …
Browse files Browse the repository at this point in the history
…value for the binary mode (pythongh-102068)

(cherry picked from commit 6f25657)

Co-authored-by: Gihwan Kim <ghkim3221@gmail.com>
  • Loading branch information
0x1306e6d authored and miss-islington committed Feb 21, 2023
1 parent 7bb41d9 commit f8aabb5
Show file tree
Hide file tree
Showing 4 changed files with 32 additions and 12 deletions.
2 changes: 1 addition & 1 deletion Lib/fileinput.py
Original file line number Diff line number Diff line change
Expand Up @@ -419,7 +419,7 @@ def isstdin(self):


def hook_compressed(filename, mode, *, encoding=None, errors=None):
if encoding is None: # EncodingWarning is emitted in FileInput() already.
if encoding is None and "b" not in mode: # EncodingWarning is emitted in FileInput() already.
encoding = "locale"
ext = os.path.splitext(filename)[1]
if ext == '.gz':
Expand Down
39 changes: 28 additions & 11 deletions Lib/test/test_fileinput.py
Original file line number Diff line number Diff line change
Expand Up @@ -903,29 +903,29 @@ def setUp(self):
self.fake_open = InvocationRecorder()

def test_empty_string(self):
self.do_test_use_builtin_open("", 1)
self.do_test_use_builtin_open_text("", "r")

def test_no_ext(self):
self.do_test_use_builtin_open("abcd", 2)
self.do_test_use_builtin_open_text("abcd", "r")

@unittest.skipUnless(gzip, "Requires gzip and zlib")
def test_gz_ext_fake(self):
original_open = gzip.open
gzip.open = self.fake_open
try:
result = fileinput.hook_compressed("test.gz", "3")
result = fileinput.hook_compressed("test.gz", "r")
finally:
gzip.open = original_open

self.assertEqual(self.fake_open.invocation_count, 1)
self.assertEqual(self.fake_open.last_invocation, (("test.gz", "3"), {}))
self.assertEqual(self.fake_open.last_invocation, (("test.gz", "r"), {}))

@unittest.skipUnless(gzip, "Requires gzip and zlib")
def test_gz_with_encoding_fake(self):
original_open = gzip.open
gzip.open = lambda filename, mode: io.BytesIO(b'Ex-binary string')
try:
result = fileinput.hook_compressed("test.gz", "3", encoding="utf-8")
result = fileinput.hook_compressed("test.gz", "r", encoding="utf-8")
finally:
gzip.open = original_open
self.assertEqual(list(result), ['Ex-binary string'])
Expand All @@ -935,23 +935,40 @@ def test_bz2_ext_fake(self):
original_open = bz2.BZ2File
bz2.BZ2File = self.fake_open
try:
result = fileinput.hook_compressed("test.bz2", "4")
result = fileinput.hook_compressed("test.bz2", "r")
finally:
bz2.BZ2File = original_open

self.assertEqual(self.fake_open.invocation_count, 1)
self.assertEqual(self.fake_open.last_invocation, (("test.bz2", "4"), {}))
self.assertEqual(self.fake_open.last_invocation, (("test.bz2", "r"), {}))

def test_blah_ext(self):
self.do_test_use_builtin_open("abcd.blah", "5")
self.do_test_use_builtin_open_binary("abcd.blah", "rb")

def test_gz_ext_builtin(self):
self.do_test_use_builtin_open("abcd.Gz", "6")
self.do_test_use_builtin_open_binary("abcd.Gz", "rb")

def test_bz2_ext_builtin(self):
self.do_test_use_builtin_open("abcd.Bz2", "7")
self.do_test_use_builtin_open_binary("abcd.Bz2", "rb")

def do_test_use_builtin_open(self, filename, mode):
def test_binary_mode_encoding(self):
self.do_test_use_builtin_open_binary("abcd", "rb")

def test_text_mode_encoding(self):
self.do_test_use_builtin_open_text("abcd", "r")

def do_test_use_builtin_open_binary(self, filename, mode):
original_open = self.replace_builtin_open(self.fake_open)
try:
result = fileinput.hook_compressed(filename, mode)
finally:
self.replace_builtin_open(original_open)

self.assertEqual(self.fake_open.invocation_count, 1)
self.assertEqual(self.fake_open.last_invocation,
((filename, mode), {'encoding': None, 'errors': None}))

def do_test_use_builtin_open_text(self, filename, mode):
original_open = self.replace_builtin_open(self.fake_open)
try:
result = fileinput.hook_compressed(filename, mode)
Expand Down
1 change: 1 addition & 0 deletions Misc/ACKS
Original file line number Diff line number Diff line change
Expand Up @@ -912,6 +912,7 @@ Tyler Kieft
Mads Kiilerich
Jason Killen
Derek D. Kim
Gihwan Kim
Jan Kim
Taek Joo Kim
Sam Kimbrel
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,2 @@
For the binary mode, :func:`fileinput.hookcompressed` doesn't set the ``encoding`` value
even if the value is ``None``. Patch by Gihwan Kim.

0 comments on commit f8aabb5

Please sign in to comment.