Issue #6135: Adds encoding and errors parameters to subprocess

Doc/library/subprocess.rst

@@ -38,7 +38,8 @@ compatibility with older versions, see the :ref:`call-function-trio` section.
 
 
 .. function:: run(args, *, stdin=None, input=None, stdout=None, stderr=None,\
-                  shell=False, timeout=None, check=False)
+                  shell=False, timeout=None, check=False, \
+                  encoding=None, errors=None)
 
    Run the command described by *args*.  Wait for command to complete, then
    return a :class:`CompletedProcess` instance.
@@ -60,15 +61,20 @@ compatibility with older versions, see the :ref:`call-function-trio` section.
 
    The *input* argument is passed to :meth:`Popen.communicate` and thus to the
    subprocess's stdin.  If used it must be a byte sequence, or a string if
-   ``universal_newlines=True``.  When used, the internal :class:`Popen` object
+   *encoding* or *errors* is specified or *universal_newlines* is True.  When
-   is automatically created with ``stdin=PIPE``, and the *stdin* argument may
+   used, the internal :class:`Popen` object is automatically created with
-   not be used as well.
+   ``stdin=PIPE``, and the *stdin* argument may not be used as well.
 
    If *check* is True, and the process exits with a non-zero exit code, a
    :exc:`CalledProcessError` exception will be raised. Attributes of that
    exception hold the arguments, the exit code, and stdout and stderr if they
    were captured.
 
+   If *encoding* or *errors* are specified, or *universal_newlines* is True,
+   file objects for stdin, stdout and stderr are opened in text mode using the
+   specified *encoding* and *errors* or the :class:`io.TextIOWrapper` default.
+   Otherwise, file objects are opened in binary mode.
+
    Examples::
 
       >>> subprocess.run(["ls", "-l"])  # doesn't capture output
@@ -85,6 +91,10 @@ compatibility with older versions, see the :ref:`call-function-trio` section.
 
    .. versionadded:: 3.5
 
+   .. versionchanged:: 3.6
+
+      Added *encoding* and *errors* parameters
+
 .. class:: CompletedProcess
 
    The return value from :func:`run`, representing a process that has finished.
@@ -104,8 +114,8 @@ compatibility with older versions, see the :ref:`call-function-trio` section.
    .. attribute:: stdout
 
       Captured stdout from the child process. A bytes sequence, or a string if
-      :func:`run` was called with ``universal_newlines=True``. None if stdout
+      :func:`run` was called with an encoding or errors. None if stdout was not
-      was not captured.
+      captured.
 
       If you ran the process with ``stderr=subprocess.STDOUT``, stdout and
       stderr will be combined in this attribute, and :attr:`stderr` will be
@@ -114,8 +124,8 @@ compatibility with older versions, see the :ref:`call-function-trio` section.
    .. attribute:: stderr
 
       Captured stderr from the child process. A bytes sequence, or a string if
-      :func:`run` was called with ``universal_newlines=True``. None if stderr
+      :func:`run` was called with an encoding or errors. None if stderr was not
-      was not captured.
+      captured.
 
    .. method:: check_returncode()
 
@@ -249,19 +259,22 @@ default values. The arguments that are most commonly needed are:
    .. index::
       single: universal newlines; subprocess module
 
-   If *universal_newlines* is ``False`` the file objects *stdin*, *stdout* and
+   If *encoding* or *errors* are specified, or *universal_newlines* is True,
-   *stderr* will be opened as binary streams, and no line ending conversion is
+   the file objects *stdin*, *stdout* and *stderr* will be opened in text
-   done.
+   mode using the *encoding* and *errors* specified in the call or the
+   defaults for :class:`io.TextIOWrapper`.
 
-   If *universal_newlines* is ``True``, these file objects
+   For *stdin*, line ending characters ``'\n'`` in the input will be converted
-   will be opened as text streams in :term:`universal newlines` mode
+   to the default line separator :data:`os.linesep`. For *stdout* and *stderr*,
-   using the encoding returned by :func:`locale.getpreferredencoding(False)
+   all line endings in the output will be converted to ``'\n'``.  For more
-   <locale.getpreferredencoding>`.  For *stdin*, line ending characters
+   information see the documentation of the :class:`io.TextIOWrapper` class
-   ``'\n'`` in the input will be converted to the default line separator
+   when the *newline* argument to its constructor is ``None``.
-   :data:`os.linesep`.  For *stdout* and *stderr*, all line endings in the
+
-   output will be converted to ``'\n'``.  For more information see the
+   If text mode is not used, *stdin*, *stdout* and *stderr* will be opened as
-   documentation of the :class:`io.TextIOWrapper` class when the *newline*
+   binary streams. No encoding or line ending conversion is performed.
-   argument to its constructor is ``None``.
+
+   .. versionadded:: 3.6
+      Added *encoding* and *errors* parameters.
 
    .. note::
 
@@ -306,7 +319,8 @@ functions.
                  stderr=None, preexec_fn=None, close_fds=True, shell=False, \
                  cwd=None, env=None, universal_newlines=False, \
                  startupinfo=None, creationflags=0, restore_signals=True, \
-                 start_new_session=False, pass_fds=())
+                 start_new_session=False, pass_fds=(), *, \
+                 encoding=None, errors=None)
 
    Execute a child program in a new process.  On POSIX, the class uses
    :meth:`os.execvp`-like behavior to execute the child program.  On Windows,
@@ -482,10 +496,14 @@ functions.
 
    .. _side-by-side assembly: https://en.wikipedia.org/wiki/Side-by-Side_Assembly
 
-   If *universal_newlines* is ``True``, the file objects *stdin*, *stdout*
+   If *encoding* or *errors* are specified, the file objects *stdin*, *stdout*
-   and *stderr* are opened as text streams in universal newlines mode, as
+   and *stderr* are opened in text mode with the specified encoding and
-   described above in :ref:`frequently-used-arguments`, otherwise they are
+   *errors*, as described above in :ref:`frequently-used-arguments`. If
-   opened as binary streams.
+   *universal_newlines* is ``True``, they are opened in text mode with default
+   encoding. Otherwise, they are opened as binary streams.
+
+   .. versionadded:: 3.6
+      *encoding* and *errors* were added.
 
    If given, *startupinfo* will be a :class:`STARTUPINFO` object, which is
    passed to the underlying ``CreateProcess`` function.
@@ -601,11 +619,12 @@ Instances of the :class:`Popen` class have the following methods:
    Interact with process: Send data to stdin.  Read data from stdout and stderr,
    until end-of-file is reached.  Wait for process to terminate.  The optional
    *input* argument should be data to be sent to the child process, or
-   ``None``, if no data should be sent to the child.  The type of *input*
+   ``None``, if no data should be sent to the child.  If streams were opened in
-   must be bytes or, if *universal_newlines* was ``True``, a string.
+   text mode, *input* must be a string.  Otherwise, it must be bytes.
 
    :meth:`communicate` returns a tuple ``(stdout_data, stderr_data)``.
-   The data will be bytes or, if *universal_newlines* was ``True``, strings.
+   The data will be strings if streams were opened in text mode; otherwise,
+   bytes.
 
    Note that if you want to send data to the process's stdin, you need to create
    the Popen object with ``stdin=PIPE``.  Similarly, to get anything other than
@@ -672,28 +691,30 @@ The following attributes are also available:
 .. attribute:: Popen.stdin
 
    If the *stdin* argument was :data:`PIPE`, this attribute is a writeable
-   stream object as returned by :func:`open`. If the *universal_newlines*
+   stream object as returned by :func:`open`. If the *encoding* or *errors*
-   argument was ``True``, the stream is a text stream, otherwise it is a byte
+   arguments were specified or the *universal_newlines* argument was ``True``,
-   stream. If the *stdin* argument was not :data:`PIPE`, this attribute is
+   the stream is a text stream, otherwise it is a byte stream. If the *stdin*
-   ``None``.
+   argument was not :data:`PIPE`, this attribute is ``None``.
 
 
 .. attribute:: Popen.stdout
 
    If the *stdout* argument was :data:`PIPE`, this attribute is a readable
    stream object as returned by :func:`open`. Reading from the stream provides
-   output from the child process. If the *universal_newlines* argument was
+   output from the child process. If the *encoding* or *errors* arguments were
-   ``True``, the stream is a text stream, otherwise it is a byte stream. If the
+   specified or the *universal_newlines* argument was ``True``, the stream is a
-   *stdout* argument was not :data:`PIPE`, this attribute is ``None``.
+   text stream, otherwise it is a byte stream. If the *stdout* argument was not
+   :data:`PIPE`, this attribute is ``None``.
 
 
 .. attribute:: Popen.stderr
 
    If the *stderr* argument was :data:`PIPE`, this attribute is a readable
    stream object as returned by :func:`open`. Reading from the stream provides
-   error output from the child process. If the *universal_newlines* argument was
+   error output from the child process. If the *encoding* or *errors* arguments
-   ``True``, the stream is a text stream, otherwise it is a byte stream. If the
+   were specified or the *universal_newlines* argument was ``True``, the stream
-   *stderr* argument was not :data:`PIPE`, this attribute is ``None``.
+   is a text stream, otherwise it is a byte stream. If the *stderr* argument was
+   not :data:`PIPE`, this attribute is ``None``.
 
 .. warning::
 
@@ -886,7 +907,9 @@ calls these functions.
       *timeout* was added.
 
 
-.. function:: check_output(args, *, stdin=None, stderr=None, shell=False, universal_newlines=False, timeout=None)
+.. function:: check_output(args, *, stdin=None, stderr=None, shell=False, \
+                           encoding=None, errors=None, \
+                           universal_newlines=False, timeout=None)
 
    Run command with arguments and return its output.
 
@@ -1142,7 +1165,7 @@ handling consistency are valid for these functions.
    Return ``(status, output)`` of executing *cmd* in a shell.
 
    Execute the string *cmd* in a shell with :meth:`Popen.check_output` and
-   return a 2-tuple ``(status, output)``. Universal newlines mode is used;
+   return a 2-tuple ``(status, output)``. The locale encoding is used;
    see the notes on :ref:`frequently-used-arguments` for more details.
 
    A trailing newline is stripped from the output.

Doc/whatsnew/3.6.rst

@@ -589,6 +589,9 @@ proc: ...``) or call explicitly the :meth:`~subprocess.Popen.wait` method to
 read the exit status of the child process (Contributed by Victor Stinner in
 :issue:`26741`).
 
+The :class:`subprocess.Popen` constructor and all functions that pass arguments
+through to it now accept *encoding* and *errors* arguments. Specifying either
+of these will enable text mode for the *stdin*, *stdout* and *stderr* streams.
 
 telnetlib
 ---------

Lib/subprocess.py

@@ -30,7 +30,8 @@ class Popen(args, bufsize=-1, executable=None,
             preexec_fn=None, close_fds=True, shell=False,
             cwd=None, env=None, universal_newlines=False,
             startupinfo=None, creationflags=0,
-            restore_signals=True, start_new_session=False, pass_fds=()):
+            restore_signals=True, start_new_session=False, pass_fds=(),
+            *, encoding=None, errors=None):
 
 
 Arguments are:
@@ -104,20 +105,13 @@ in the child process prior to executing the command.
 If env is not None, it defines the environment variables for the new
 process.
 
-If universal_newlines is False, the file objects stdin, stdout and stderr
+If encoding or errors are specified or universal_newlines is True, the file
-are opened as binary files, and no line ending conversion is done.
+objects stdout and stderr are opened in text mode. See io.TextIOWrapper for
+the interpretation of these parameters are used.
 
-If universal_newlines is True, the file objects stdout and stderr are
+If no encoding is specified and universal_newlines is False, the file
-opened as a text file, but lines may be terminated by any of '\n',
+objects stdin, stdout and stderr are opened as binary files, and no
-the Unix end-of-line convention, '\r', the old Macintosh convention or
+line ending conversion is done.
-'\r\n', the Windows convention.  All of these external representations
-are seen as '\n' by the Python program.  Also, the newlines attribute
-of the file objects stdout, stdin and stderr are not updated by the
-communicate() method.
-
-In either case, the process being communicated with should start up
-expecting to receive bytes on its standard input and decode them with
-the same encoding they are sent in.
 
 The startupinfo and creationflags, if given, will be passed to the
 underlying CreateProcess() function.  They can specify things such as
@@ -234,11 +228,8 @@ communicate(input=None)
     and stderr, until end-of-file is reached.  Wait for process to
     terminate.  The optional input argument should be data to be
     sent to the child process, or None, if no data should be sent to
-    the child. If the Popen instance was constructed with universal_newlines
+    the child. If the Popen instance was constructed in text mode, the
-    set to True, the input argument should be a string and will be encoded
+    input argument should be a string. Otherwise, it should be bytes.
-    using the preferred system encoding (see locale.getpreferredencoding);
-    if universal_newlines is False, the input argument should be a
-    byte string.
 
     communicate() returns a tuple (stdout, stderr).
 
@@ -808,8 +799,8 @@ def getstatusoutput(cmd):
     """    Return (status, output) of executing cmd in a shell.
 
     Execute the string 'cmd' in a shell with 'check_output' and
-    return a 2-tuple (status, output). Universal newlines mode is used,
+    return a 2-tuple (status, output). The locale encoding is used
-    meaning that the result with be decoded to a string.
+    to decode the output and process newlines.
 
     A trailing newline is stripped from the output.
     The exit status for the command can be interpreted
@@ -859,7 +850,7 @@ class Popen(object):
                  shell=False, cwd=None, env=None, universal_newlines=False,
                  startupinfo=None, creationflags=0,
                  restore_signals=True, start_new_session=False,
-                 pass_fds=()):
+                 pass_fds=(), *, encoding=None, errors=None):
         """Create new Popen instance."""
         _cleanup()
         # Held while anything is calling waitpid before returncode has been
@@ -912,6 +903,8 @@ class Popen(object):
         self.pid = None
         self.returncode = None
         self.universal_newlines = universal_newlines
+        self.encoding = encoding
+        self.errors = errors
 
         # Input and output objects. The general principle is like
         # this:
@@ -944,22 +937,28 @@ class Popen(object):
             if errread != -1:
                 errread = msvcrt.open_osfhandle(errread.Detach(), 0)
 
-        if p2cwrite != -1:
+        text_mode = encoding or errors or universal_newlines
-            self.stdin = io.open(p2cwrite, 'wb', bufsize)
-            if universal_newlines:
-                self.stdin = io.TextIOWrapper(self.stdin, write_through=True,
-                                              line_buffering=(bufsize == 1))
-        if c2pread != -1:
-            self.stdout = io.open(c2pread, 'rb', bufsize)
-            if universal_newlines:
-                self.stdout = io.TextIOWrapper(self.stdout)
-        if errread != -1:
-            self.stderr = io.open(errread, 'rb', bufsize)
-            if universal_newlines:
-                self.stderr = io.TextIOWrapper(self.stderr)
 
         self._closed_child_pipe_fds = False
+
         try:
+            if p2cwrite != -1:
+                self.stdin = io.open(p2cwrite, 'wb', bufsize)
+                if text_mode:
+                    self.stdin = io.TextIOWrapper(self.stdin, write_through=True,
+                            line_buffering=(bufsize == 1),
+                            encoding=encoding, errors=errors)
+            if c2pread != -1:
+                self.stdout = io.open(c2pread, 'rb', bufsize)
+                if text_mode:
+                    self.stdout = io.TextIOWrapper(self.stdout,
+                            encoding=encoding, errors=errors)
+            if errread != -1:
+                self.stderr = io.open(errread, 'rb', bufsize)
+                if text_mode:
+                    self.stderr = io.TextIOWrapper(self.stderr,
+                            encoding=encoding, errors=errors)
+
             self._execute_child(args, executable, preexec_fn, close_fds,
                                 pass_fds, cwd, env,
                                 startupinfo, creationflags, shell,
@@ -993,8 +992,8 @@ class Popen(object):
 
             raise
 
-    def _translate_newlines(self, data, encoding):
+    def _translate_newlines(self, data, encoding, errors):
-        data = data.decode(encoding)
+        data = data.decode(encoding, errors)
         return data.replace("\r\n", "\n").replace("\r", "\n")
 
     def __enter__(self):
@@ -1779,13 +1778,15 @@ class Popen(object):
 
             # Translate newlines, if requested.
             # This also turns bytes into strings.
-            if self.universal_newlines:
+            if self.encoding or self.errors or self.universal_newlines:
                 if stdout is not None:
                     stdout = self._translate_newlines(stdout,
-                                                      self.stdout.encoding)
+                                                      self.stdout.encoding,
+                                                      self.stdout.errors)
                 if stderr is not None:
                     stderr = self._translate_newlines(stderr,
-                                                      self.stderr.encoding)
+                                                      self.stderr.encoding,
+                                                      self.stderr.errors)
 
             return (stdout, stderr)
 
@@ -1797,8 +1798,10 @@ class Popen(object):
             if self.stdin and self._input is None:
                 self._input_offset = 0
                 self._input = input
-                if self.universal_newlines and input is not None:
+                if input is not None and (
-                    self._input = self._input.encode(self.stdin.encoding)
+                    self.encoding or self.errors or self.universal_newlines):
+                    self._input = self._input.encode(self.stdin.encoding,
+                                                     self.stdin.errors)
 
 
         def send_signal(self, sig):

Lib/test/test_subprocess.py

@@ -894,31 +894,42 @@ class ProcessTestCase(BaseTestCase):
         #
         # UTF-16 and UTF-32-BE are sufficient to check both with BOM and
         # without, and UTF-16 and UTF-32.
-        import _bootlocale
         for encoding in ['utf-16', 'utf-32-be']:
-            old_getpreferredencoding = _bootlocale.getpreferredencoding
-            # Indirectly via io.TextIOWrapper, Popen() defaults to
-            # locale.getpreferredencoding(False) and earlier in Python 3.2 to
-            # locale.getpreferredencoding().
-            def getpreferredencoding(do_setlocale=True):
-                return encoding
             code = ("import sys; "
                     r"sys.stdout.buffer.write('1\r\n2\r3\n4'.encode('%s'))" %
                     encoding)
             args = [sys.executable, '-c', code]
-            try:
-                _bootlocale.getpreferredencoding = getpreferredencoding
             # We set stdin to be non-None because, as of this writing,
             # a different code path is used when the number of pipes is
             # zero or one.
-                popen = subprocess.Popen(args, universal_newlines=True,
+            popen = subprocess.Popen(args,
                                      stdin=subprocess.PIPE,
-                                         stdout=subprocess.PIPE)
+                                     stdout=subprocess.PIPE,
+                                     encoding=encoding)
             stdout, stderr = popen.communicate(input='')
-            finally:
-                _bootlocale.getpreferredencoding = old_getpreferredencoding
             self.assertEqual(stdout, '1\n2\n3\n4')
 
+    def test_communicate_errors(self):
+        for errors, expected in [
+            ('ignore', ''),
+            ('replace', '\ufffd\ufffd'),
+            ('surrogateescape', '\udc80\udc80'),
+            ('backslashreplace', '\\x80\\x80'),
+        ]:
+            code = ("import sys; "
+                    r"sys.stdout.buffer.write(b'[\x80\x80]')")
+            args = [sys.executable, '-c', code]
+            # We set stdin to be non-None because, as of this writing,
+            # a different code path is used when the number of pipes is
+            # zero or one.
+            popen = subprocess.Popen(args,
+                                     stdin=subprocess.PIPE,
+                                     stdout=subprocess.PIPE,
+                                     encoding='utf-8',
+                                     errors=errors)
+            stdout, stderr = popen.communicate(input='')
+            self.assertEqual(stdout, '[{}]'.format(expected))
+
     def test_no_leaking(self):
         # Make sure we leak no resources
         if not mswindows:
@@ -2539,6 +2550,18 @@ class Win32ProcessTestCase(BaseTestCase):
         with p:
             self.assertIn(b"physalis", p.stdout.read())
 
+    def test_shell_encodings(self):
+        # Run command through the shell (string)
+        for enc in ['ansi', 'oem']:
+            newenv = os.environ.copy()
+            newenv["FRUIT"] = "physalis"
+            p = subprocess.Popen("set", shell=1,
+                                 stdout=subprocess.PIPE,
+                                 env=newenv,
+                                 encoding=enc)
+            with p:
+                self.assertIn("physalis", p.stdout.read(), enc)
+
     def test_call_string(self):
         # call() function with string argument on Windows
         rc = subprocess.call(sys.executable +

Misc/NEWS

@@ -237,6 +237,8 @@ Build
 Windows
 -------
 
+- Issue #6135: Adds encoding and errors parameters to subprocess.
+
 - Issue #27959: Adds oem encoding, alias ansi to mbcs, move aliasmbcs to
   codec lookup.