[3.11] gh-101144: Allow zipfile.Path .open & .read_text encoding to be positional (#101179)

The zipfile.Path open() and read_text() encoding parameter can be supplied as a positional argument without causing a TypeError again. 3.10.0b1 included a regression that made it keyword only.

Documentation update included as users writing code to be compatible with a wide range of versions will need to consider this for some time..

(cherry picked from commit 5927013e47)

Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google]

Doc/library/zipfile.rst

@@ -551,6 +551,12 @@ Path objects are traversable using the ``/`` operator or ``joinpath``.
       Added support for text and binary modes for open. Default
       mode is now text.
 
+   .. versionchanged:: 3.11.2
+      The ``encoding`` parameter can be supplied as a positional argument
+      without causing a :exc:`TypeError`. As it could in 3.9. Code needing to
+      be compatible with unpatched 3.10 and 3.11 versions must pass all
+      :class:`io.TextIOWrapper` arguments, ``encoding`` included, as keywords.
+
 .. method:: Path.iterdir()
 
    Enumerate the children of the current directory.
@@ -596,6 +602,12 @@ Path objects are traversable using the ``/`` operator or ``joinpath``.
    :class:`io.TextIOWrapper` (except ``buffer``, which is
    implied by the context).
 
+   .. versionchanged:: 3.11.2
+      The ``encoding`` parameter can be supplied as a positional argument
+      without causing a :exc:`TypeError`. As it could in 3.9. Code needing to
+      be compatible with unpatched 3.10 and 3.11 versions must pass all
+      :class:`io.TextIOWrapper` arguments, ``encoding`` included, as keywords.
+
 .. method:: Path.read_bytes()
 
    Read the current file as bytes.

Lib/test/test_zipfile.py

@@ -10,6 +10,7 @@ import string
 import struct
 import subprocess
 import sys
+from test.support.script_helper import assert_python_ok
 import time
 import unittest
 import unittest.mock as mock
@@ -3005,7 +3006,69 @@ class TestPath(unittest.TestCase):
         a, b, g = root.iterdir()
         with a.open(encoding="utf-8") as strm:
             data = strm.read()
-        assert data == "content of a"
+        self.assertEqual(data, "content of a")
+        with a.open('r', "utf-8") as strm:  # not a kw, no gh-101144 TypeError
+            data = strm.read()
+        self.assertEqual(data, "content of a")
+
+    def test_open_encoding_utf16(self):
+        in_memory_file = io.BytesIO()
+        zf = zipfile.ZipFile(in_memory_file, "w")
+        zf.writestr("path/16.txt", "This was utf-16".encode("utf-16"))
+        zf.filename = "test_open_utf16.zip"
+        root = zipfile.Path(zf)
+        (path,) = root.iterdir()
+        u16 = path.joinpath("16.txt")
+        with u16.open('r', "utf-16") as strm:
+            data = strm.read()
+        self.assertEqual(data, "This was utf-16")
+        with u16.open(encoding="utf-16") as strm:
+            data = strm.read()
+        self.assertEqual(data, "This was utf-16")
+
+    def test_open_encoding_errors(self):
+        in_memory_file = io.BytesIO()
+        zf = zipfile.ZipFile(in_memory_file, "w")
+        zf.writestr("path/bad-utf8.bin", b"invalid utf-8: \xff\xff.")
+        zf.filename = "test_read_text_encoding_errors.zip"
+        root = zipfile.Path(zf)
+        (path,) = root.iterdir()
+        u16 = path.joinpath("bad-utf8.bin")
+
+        # encoding= as a positional argument for gh-101144.
+        data = u16.read_text("utf-8", errors="ignore")
+        self.assertEqual(data, "invalid utf-8: .")
+        with u16.open("r", "utf-8", errors="surrogateescape") as f:
+            self.assertEqual(f.read(), "invalid utf-8: \udcff\udcff.")
+
+        # encoding= both positional and keyword is an error; gh-101144.
+        with self.assertRaisesRegex(TypeError, "encoding"):
+            data = u16.read_text("utf-8", encoding="utf-8")
+
+        # both keyword arguments work.
+        with u16.open("r", encoding="utf-8", errors="strict") as f:
+            # error during decoding with wrong codec.
+            with self.assertRaises(UnicodeDecodeError):
+                f.read()
+
+    def test_encoding_warnings(self):
+        """EncodingWarning must blame the read_text and open calls."""
+        code = '''\
+import io, zipfile
+with zipfile.ZipFile(io.BytesIO(), "w") as zf:
+    zf.filename = '<test_encoding_warnings in memory zip file>'
+    zf.writestr("path/file.txt", b"Spanish Inquisition")
+    root = zipfile.Path(zf)
+    (path,) = root.iterdir()
+    file_path = path.joinpath("file.txt")
+    unused = file_path.read_text()  # should warn
+    file_path.open("r").close()  # should warn
+'''
+        proc = assert_python_ok('-X', 'warn_default_encoding', '-c', code)
+        warnings = proc.err.splitlines()
+        self.assertEqual(len(warnings), 2, proc.err)
+        self.assertRegex(warnings[0], rb"^<string>:8: EncodingWarning:")
+        self.assertRegex(warnings[1], rb"^<string>:9: EncodingWarning:")
 
     def test_open_write(self):
         """
@@ -3047,6 +3110,7 @@ class TestPath(unittest.TestCase):
         root = zipfile.Path(alpharep)
         a, b, g = root.iterdir()
         assert a.read_text(encoding="utf-8") == "content of a"
+        a.read_text("utf-8")  # No positional arg TypeError per gh-101144.
         assert a.read_bytes() == b"content of a"
 
     @pass_alpharep

Lib/zipfile.py

@@ -2287,6 +2287,11 @@ class FastLookup(CompleteDirs):
         return self.__lookup
 
 
+def _extract_text_encoding(encoding=None, *args, **kwargs):
+    # stacklevel=3 so that the caller of the caller see any warning.
+    return io.text_encoding(encoding, 3), args, kwargs
+
+
 class Path:
     """
     A pathlib-compatible interface for zip files.
@@ -2396,9 +2401,9 @@ class Path:
             if args or kwargs:
                 raise ValueError("encoding args invalid for binary operation")
             return stream
-        else:
-            kwargs["encoding"] = io.text_encoding(kwargs.get("encoding"))
-        return io.TextIOWrapper(stream, *args, **kwargs)
+        # Text mode:
+        encoding, args, kwargs = _extract_text_encoding(*args, **kwargs)
+        return io.TextIOWrapper(stream, encoding, *args, **kwargs)
 
     @property
     def name(self):
@@ -2421,8 +2426,8 @@ class Path:
         return pathlib.Path(self.root.filename).joinpath(self.at)
 
     def read_text(self, *args, **kwargs):
-        kwargs["encoding"] = io.text_encoding(kwargs.get("encoding"))
-        with self.open('r', *args, **kwargs) as strm:
+        encoding, args, kwargs = _extract_text_encoding(*args, **kwargs)
+        with self.open('r', encoding, *args, **kwargs) as strm:
             return strm.read()
 
     def read_bytes(self):

Misc/NEWS.d/next/Library/2023-01-18-17-58-50.gh-issue-101144.FHd8Un.rst

@@ -0,0 +1,4 @@
+Make :func:`zipfile.Path.open` and :func:`zipfile.Path.read_text` also accept
+``encoding`` as a positional argument. This was the behavior in Python 3.9 and
+earlier.  3.10 introduced a regression where supplying it as a positional
+argument would lead to a :exc:`TypeError`.