Lib/nntplib.py

"""An NNTP client class based on:
- RFC 977: Network News Transfer Protocol
- RFC 2980: Common NNTP Extensions
- RFC 3977: Network News Transfer Protocol (version 2)

Example:

>>> from nntplib import NNTP
>>> s = NNTP('news')
>>> resp, count, first, last, name = s.group('comp.lang.python')
>>> print('Group', name, 'has', count, 'articles, range', first, 'to', last)
Group comp.lang.python has 51 articles, range 5770 to 5821
>>> resp, subs = s.xhdr('subject', '{0}-{1}'.format(first, last))
>>> resp = s.quit()
>>>

Here 'resp' is the server response line.
Error responses are turned into exceptions.

To post an article from a file:
>>> f = open(filename, 'rb') # file containing article, including header
>>> resp = s.post(f)
>>>

For descriptions of all methods, read the comments in the code below.
Note that all arguments and return values representing article numbers
are strings, not numbers, since they are rarely used for calculations.
"""

# RFC 977 by Brian Kantor and Phil Lapsley.
# xover, xgtitle, xpath, date methods by Kevan Heydon

# Incompatible changes from the 2.x nntplib:
# - all commands are encoded as UTF-8 data (using the "surrogateescape"
#   error handler), except for raw message data (POST, IHAVE)
# - all responses are decoded as UTF-8 data (using the "surrogateescape"
#   error handler), except for raw message data (ARTICLE, HEAD, BODY)
# - the `file` argument to various methods is keyword-only
#
# - NNTP.date() returns a datetime object
# - NNTP.newgroups() and NNTP.newnews() take a datetime (or date) object,
#   rather than a pair of (date, time) strings.
# - NNTP.newgroups() and NNTP.list() return a list of GroupInfo named tuples
# - NNTP.descriptions() returns a dict mapping group names to descriptions
# - NNTP.xover() returns a list of dicts mapping field names (header or metadata)
#   to field values; each dict representing a message overview.
# - NNTP.article(), NNTP.head() and NNTP.body() return a (response, ArticleInfo)
#   tuple.
# - the "internal" methods have been marked private (they now start with
#   an underscore)

# Other changes from the 2.x/3.1 nntplib:
# - automatic querying of capabilities at connect
# - New method NNTP.getcapabilities()
# - New method NNTP.over()
# - New helper function decode_header()
# - NNTP.post() and NNTP.ihave() accept file objects, bytes-like objects and
#   arbitrary iterables yielding lines.
# - An extensive test suite :-)

# TODO:
# - return structured data (GroupInfo etc.) everywhere
# - support HDR

# Imports
import re
import socket
import collections
import datetime
import sys

try:
    import ssl
except ImportError:
    _have_ssl = False
else:
    _have_ssl = True

from email.header import decode_header as _email_decode_header
from socket import _GLOBAL_DEFAULT_TIMEOUT

__all__ = ["NNTP",
           "NNTPError", "NNTPReplyError", "NNTPTemporaryError",
           "NNTPPermanentError", "NNTPProtocolError", "NNTPDataError",
           "decode_header",
           ]

# maximal line length when calling readline(). This is to prevent
# reading arbitrary length lines. RFC 3977 limits NNTP line length to
# 512 characters, including CRLF. We have selected 2048 just to be on
# the safe side.
_MAXLINE = 2048


# Exceptions raised when an error or invalid response is received
class NNTPError(Exception):
    """Base class for all nntplib exceptions"""
    def __init__(self, *args):
        Exception.__init__(self, *args)
        try:
            self.response = args[0]
        except IndexError:
            self.response = 'No response given'

class NNTPReplyError(NNTPError):
    """Unexpected [123]xx reply"""
    pass

class NNTPTemporaryError(NNTPError):
    """4xx errors"""
    pass

class NNTPPermanentError(NNTPError):
    """5xx errors"""
    pass

class NNTPProtocolError(NNTPError):
    """Response does not begin with [1-5]"""
    pass

class NNTPDataError(NNTPError):
    """Error in response data"""
    pass


# Standard port used by NNTP servers
NNTP_PORT = 119
NNTP_SSL_PORT = 563

# Response numbers that are followed by additional text (e.g. article)
_LONGRESP = {
    '100',   # HELP
    '101',   # CAPABILITIES
    '211',   # LISTGROUP   (also not multi-line with GROUP)
    '215',   # LIST
    '220',   # ARTICLE
    '221',   # HEAD, XHDR
    '222',   # BODY
    '224',   # OVER, XOVER
    '225',   # HDR
    '230',   # NEWNEWS
    '231',   # NEWGROUPS
    '282',   # XGTITLE
}

# Default decoded value for LIST OVERVIEW.FMT if not supported
_DEFAULT_OVERVIEW_FMT = [
    "subject", "from", "date", "message-id", "references", ":bytes", ":lines"]

# Alternative names allowed in LIST OVERVIEW.FMT response
_OVERVIEW_FMT_ALTERNATIVES = {
    'bytes': ':bytes',
    'lines': ':lines',
}

# Line terminators (we always output CRLF, but accept any of CRLF, CR, LF)
_CRLF = b'\r\n'

GroupInfo = collections.namedtuple('GroupInfo',
                                   ['group', 'last', 'first', 'flag'])

ArticleInfo = collections.namedtuple('ArticleInfo',
                                     ['number', 'message_id', 'lines'])


# Helper function(s)
def decode_header(header_str):
    """Takes a unicode string representing a munged header value
    and decodes it as a (possibly non-ASCII) readable value."""
    parts = []
    for v, enc in _email_decode_header(header_str):
        if isinstance(v, bytes):
            parts.append(v.decode(enc or 'ascii'))
        else:
            parts.append(v)
    return ''.join(parts)

def _parse_overview_fmt(lines):
    """Parse a list of string representing the response to LIST OVERVIEW.FMT
    and return a list of header/metadata names.
    Raises NNTPDataError if the response is not compliant
    (cf. RFC 3977, section 8.4)."""
    fmt = []
    for line in lines:
        if line[0] == ':':
            # Metadata name (e.g. ":bytes")
            name, _, suffix = line[1:].partition(':')
            name = ':' + name
        else:
            # Header name (e.g. "Subject:" or "Xref:full")
            name, _, suffix = line.partition(':')
        name = name.lower()
        name = _OVERVIEW_FMT_ALTERNATIVES.get(name, name)
        # Should we do something with the suffix?
        fmt.append(name)
    defaults = _DEFAULT_OVERVIEW_FMT
    if len(fmt) < len(defaults):
        raise NNTPDataError("LIST OVERVIEW.FMT response too short")
    if fmt[:len(defaults)] != defaults:
        raise NNTPDataError("LIST OVERVIEW.FMT redefines default fields")
    return fmt

def _parse_overview(lines, fmt, data_process_func=None):
    """Parse the response to an OVER or XOVER command according to the
    overview format `fmt`."""
    n_defaults = len(_DEFAULT_OVERVIEW_FMT)
    overview = []
    for line in lines:
        fields = {}
        article_number, *tokens = line.split('\t')
        article_number = int(article_number)
        for i, token in enumerate(tokens):
            if i >= len(fmt):
                # XXX should we raise an error? Some servers might not
                # support LIST OVERVIEW.FMT and still return additional
                # headers.
                continue
            field_name = fmt[i]
            is_metadata = field_name.startswith(':')
            if i >= n_defaults and not is_metadata:
                # Non-default header names are included in full in the response
                # (unless the field is totally empty)
                h = field_name + ": "
                if token and token[:len(h)].lower() != h:
                    raise NNTPDataError("OVER/XOVER response doesn't include "
                                        "names of additional headers")
                token = token[len(h):] if token else None
            fields[fmt[i]] = token
        overview.append((article_number, fields))
    return overview

def _parse_datetime(date_str, time_str=None):
    """Parse a pair of (date, time) strings, and return a datetime object.
    If only the date is given, it is assumed to be date and time
    concatenated together (e.g. response to the DATE command).
    """
    if time_str is None:
        time_str = date_str[-6:]
        date_str = date_str[:-6]
    hours = int(time_str[:2])
    minutes = int(time_str[2:4])
    seconds = int(time_str[4:])
    year = int(date_str[:-4])
    month = int(date_str[-4:-2])
    day = int(date_str[-2:])
    # RFC 3977 doesn't say how to interpret 2-char years.  Assume that
    # there are no dates before 1970 on Usenet.
    if year < 70:
        year += 2000
    elif year < 100:
        year += 1900
    return datetime.datetime(year, month, day, hours, minutes, seconds)

def _unparse_datetime(dt, legacy=False):
    """Format a date or datetime object as a pair of (date, time) strings
    in the format required by the NEWNEWS and NEWGROUPS commands.  If a
    date object is passed, the time is assumed to be midnight (00h00).

    The returned representation depends on the legacy flag:
    * if legacy is False (the default):
      date has the YYYYMMDD format and time the HHMMSS format
    * if legacy is True:
      date has the YYMMDD format and time the HHMMSS format.
    RFC 3977 compliant servers should understand both formats; therefore,
    legacy is only needed when talking to old servers.
    """
    if not isinstance(dt, datetime.datetime):
        time_str = "000000"
    else:
        time_str = "{0.hour:02d}{0.minute:02d}{0.second:02d}".format(dt)
    y = dt.year
    if legacy:
        y = y % 100
        date_str = "{0:02d}{1.month:02d}{1.day:02d}".format(y, dt)
    else:
        date_str = "{0:04d}{1.month:02d}{1.day:02d}".format(y, dt)
    return date_str, time_str


if _have_ssl:

    def _encrypt_on(sock, context, hostname):
        """Wrap a socket in SSL/TLS. Arguments:
        - sock: Socket to wrap
        - context: SSL context to use for the encrypted connection
        Returns:
        - sock: New, encrypted socket.
        """
        # Generate a default SSL context if none was passed.
        if context is None:
            context = ssl._create_stdlib_context()
        return context.wrap_socket(sock, server_hostname=hostname)


# The classes themselves
class _NNTPBase:
    # UTF-8 is the character set for all NNTP commands and responses: they
    # are automatically encoded (when sending) and decoded (and receiving)
    # by this class.
    # However, some multi-line data blocks can contain arbitrary bytes (for
    # example, latin-1 or utf-16 data in the body of a message). Commands
    # taking (POST, IHAVE) or returning (HEAD, BODY, ARTICLE) raw message
    # data will therefore only accept and produce bytes objects.
    # Furthermore, since there could be non-compliant servers out there,
    # we use 'surrogateescape' as the error handler for fault tolerance
    # and easy round-tripping. This could be useful for some applications
    # (e.g. NNTP gateways).

    encoding = 'utf-8'
    errors = 'surrogateescape'

    def __init__(self, file, host,
                 readermode=None, timeout=_GLOBAL_DEFAULT_TIMEOUT):
        """Initialize an instance.  Arguments:
        - file: file-like object (open for read/write in binary mode)
        - host: hostname of the server
        - readermode: if true, send 'mode reader' command after
                      connecting.
        - timeout: timeout (in seconds) used for socket connections

        readermode is sometimes necessary if you are connecting to an
        NNTP server on the local machine and intend to call
        reader-specific commands, such as `group'.  If you get
        unexpected NNTPPermanentErrors, you might need to set
        readermode.
        """
        self.host = host
        self.file = file
        self.debugging = 0
        self.welcome = self._getresp()

        # Inquire about capabilities (RFC 3977).
        self._caps = None
        self.getcapabilities()

        # 'MODE READER' is sometimes necessary to enable 'reader' mode.
        # However, the order in which 'MODE READER' and 'AUTHINFO' need to
        # arrive differs between some NNTP servers. If _setreadermode() fails
        # with an authorization failed error, it will set this to True;
        # the login() routine will interpret that as a request to try again
        # after performing its normal function.
        # Enable only if we're not already in READER mode anyway.
        self.readermode_afterauth = False
        if readermode and 'READER' not in self._caps:
            self._setreadermode()
            if not self.readermode_afterauth:
                # Capabilities might have changed after MODE READER
                self._caps = None
                self.getcapabilities()

        # RFC 4642 2.2.2: Both the client and the server MUST know if there is
        # a TLS session active.  A client MUST NOT attempt to start a TLS
        # session if a TLS session is already active.
        self.tls_on = False

        # Log in and encryption setup order is left to subclasses.
        self.authenticated = False

    def __enter__(self):
        return self

    def __exit__(self, *args):
        is_connected = lambda: hasattr(self, "file")
        if is_connected():
            try:
                self.quit()
            except (OSError, EOFError):
                pass
            finally:
                if is_connected():
                    self._close()

    def getwelcome(self):
        """Get the welcome message from the server
        (this is read and squirreled away by __init__()).
        If the response code is 200, posting is allowed;
        if it 201, posting is not allowed."""

        if self.debugging: print('*welcome*', repr(self.welcome))
        return self.welcome

    def getcapabilities(self):
        """Get the server capabilities, as read by __init__().
        If the CAPABILITIES command is not supported, an empty dict is
        returned."""
        if self._caps is None:
            self.nntp_version = 1
            self.nntp_implementation = None
            try:
                resp, caps = self.capabilities()
            except (NNTPPermanentError, NNTPTemporaryError):
                # Server doesn't support capabilities
                self._caps = {}
            else:
                self._caps = caps
                if 'VERSION' in caps:
                    # The server can advertise several supported versions,
                    # choose the highest.
                    self.nntp_version = max(map(int, caps['VERSION']))
                if 'IMPLEMENTATION' in caps:
                    self.nntp_implementation = ' '.join(caps['IMPLEMENTATION'])
        return self._caps

    def set_debuglevel(self, level):
        """Set the debugging level.  Argument 'level' means:
        0: no debugging output (default)
        1: print commands and responses but not body text etc.
        2: also print raw lines read and sent before stripping CR/LF"""

        self.debugging = level
    debug = set_debuglevel

    def _putline(self, line):
        """Internal: send one line to the server, appending CRLF.
        The `line` must be a bytes-like object."""
        sys.audit("nntplib.putline", self, line)
        line = line + _CRLF
        if self.debugging > 1: print('*put*', repr(line))
        self.file.write(line)
        self.file.flush()

    def _putcmd(self, line):
        """Internal: send one command to the server (through _putline()).
        The `line` must be a unicode string."""
        if self.debugging: print('*cmd*', repr(line))
        line = line.encode(self.encoding, self.errors)
        self._putline(line)

    def _getline(self, strip_crlf=True):
        """Internal: return one line from the server, stripping _CRLF.
        Raise EOFError if the connection is closed.
        Returns a bytes object."""
        line = self.file.readline(_MAXLINE +1)
        if len(line) > _MAXLINE:
            raise NNTPDataError('line too long')
        if self.debugging > 1:
            print('*get*', repr(line))
        if not line: raise EOFError
        if strip_crlf:
            if line[-2:] == _CRLF:
                line = line[:-2]
            elif line[-1:] in _CRLF:
                line = line[:-1]
        return line

    def _getresp(self):
        """Internal: get a response from the server.
        Raise various errors if the response indicates an error.
        Returns a unicode string."""
        resp = self._getline()
        if self.debugging: print('*resp*', repr(resp))
        resp = resp.decode(self.encoding, self.errors)
        c = resp[:1]
        if c == '4':
            raise NNTPTemporaryError(resp)
        if c == '5':
            raise NNTPPermanentError(resp)
        if c not in '123':
            raise NNTPProtocolError(resp)
        return resp

    def _getlongresp(self, file=None):
        """Internal: get a response plus following text from the server.
        Raise various errors if the response indicates an error.

        Returns a (response, lines) tuple where `response` is a unicode
        string and `lines` is a list of bytes objects.
        If `file` is a file-like object, it must be open in binary mode.
        """

        openedFile = None
        try:
            # If a string was passed then open a file with that name
            if isinstance(file, (str, bytes)):
                openedFile = file = open(file, "wb")

            resp = self._getresp()
            if resp[:3] not in _LONGRESP:
                raise NNTPReplyError(resp)

            lines = []
            if file is not None:
                # XXX lines = None instead?
                terminators = (b'.' + _CRLF, b'.\n')
                while 1:
                    line = self._getline(False)
                    if line in terminators:
                        break
                    if line.startswith(b'..'):
                        line = line[1:]
                    file.write(line)
            else:
                terminator = b'.'
                while 1:
                    line = self._getline()
                    if line == terminator:
                        break
                    if line.startswith(b'..'):
                        line = line[1:]
                    lines.append(line)
        finally:
            # If this method created the file, then it must close it
            if openedFile:
                openedFile.close()

        return resp, lines

    def _shortcmd(self, line):
        """Internal: send a command and get the response.
        Same return value as _getresp()."""
        self._putcmd(line)
        return self._getresp()

    def _longcmd(self, line, file=None):
        """Internal: send a command and get the response plus following text.
        Same return value as _getlongresp()."""
        self._putcmd(line)
        return self._getlongresp(file)

    def _longcmdstring(self, line, file=None):
        """Internal: send a command and get the response plus following text.
        Same as _longcmd() and _getlongresp(), except that the returned `lines`
        are unicode strings rather than bytes objects.
        """
        self._putcmd(line)
        resp, list = self._getlongresp(file)
        return resp, [line.decode(self.encoding, self.errors)
                      for line in list]

    def _getoverviewfmt(self):
        """Internal: get the overview format. Queries the server if not
        already done, else returns the cached value."""
        try:
            return self._cachedoverviewfmt
        except AttributeError:
            pass
        try:
            resp, lines = self._longcmdstring("LIST OVERVIEW.FMT")
        except NNTPPermanentError:
            # Not supported by server?
            fmt = _DEFAULT_OVERVIEW_FMT[:]
        else:
            fmt = _parse_overview_fmt(lines)
        self._cachedoverviewfmt = fmt
        return fmt

    def _grouplist(self, lines):
        # Parse lines into "group last first flag"
        return [GroupInfo(*line.split()) for line in lines]

    def capabilities(self):
        """Process a CAPABILITIES command.  Not supported by all servers.
        Return:
        - resp: server response if successful
        - caps: a dictionary mapping capability names to lists of tokens
        (for example {'VERSION': ['2'], 'OVER': [], LIST: ['ACTIVE', 'HEADERS'] })
        """
        caps = {}
        resp, lines = self._longcmdstring("CAPABILITIES")
        for line in lines:
            name, *tokens = line.split()
            caps[name] = tokens
        return resp, caps

    def newgroups(self, date, *, file=None):
        """Process a NEWGROUPS command.  Arguments:
        - date: a date or datetime object
        Return:
        - resp: server response if successful
        - list: list of newsgroup names
        """
        if not isinstance(date, (datetime.date, datetime.date)):
            raise TypeError(
                "the date parameter must be a date or datetime object, "
                "not '{:40}'".format(date.__class__.__name__))
        date_str, time_str = _unparse_datetime(date, self.nntp_version < 2)
        cmd = 'NEWGROUPS {0} {1}'.format(date_str, time_str)
        resp, lines = self._longcmdstring(cmd, file)
        return resp, self._grouplist(lines)

    def newnews(self, group, date, *, file=None):
        """Process a NEWNEWS command.  Arguments:
        - group: group name or '*'
        - date: a date or datetime object
        Return:
        - resp: server response if successful
        - list: list of message ids
        """
        if not isinstance(date, (datetime.date, datetime.date)):
            raise TypeError(
                "the date parameter must be a date or datetime object, "
                "not '{:40}'".format(date.__class__.__name__))
        date_str, time_str = _unparse_datetime(date, self.nntp_version < 2)
        cmd = 'NEWNEWS {0} {1} {2}'.format(group, date_str, time_str)
        return self._longcmdstring(cmd, file)

    def list(self, group_pattern=None, *, file=None):
        """Process a LIST or LIST ACTIVE command. Arguments:
        - group_pattern: a pattern indicating which groups to query
        - file: Filename string or file object to store the result in
        Returns:
        - resp: server response if successful
        - list: list of (group, last, first, flag) (strings)
        """
        if group_pattern is not None:
            command = 'LIST ACTIVE ' + group_pattern
        else:
            command = 'LIST'
        resp, lines = self._longcmdstring(command, file)
        return resp, self._grouplist(lines)

    def _getdescriptions(self, group_pattern, return_all):
        line_pat = re.compile('^(?P<group>[^ \t]+)[ \t]+(.*)$')
        # Try the more std (acc. to RFC2980) LIST NEWSGROUPS first
        resp, lines = self._longcmdstring('LIST NEWSGROUPS ' + group_pattern)
        if not resp.startswith('215'):
            # Now the deprecated XGTITLE.  This either raises an error
            # or succeeds with the same output structure as LIST
            # NEWSGROUPS.
            resp, lines = self._longcmdstring('XGTITLE ' + group_pattern)
        groups = {}
        for raw_line in lines:
            match = line_pat.search(raw_line.strip())
            if match:
                name, desc = match.group(1, 2)
                if not return_all:
                    return desc
                groups[name] = desc
        if return_all:
            return resp, groups
        else:
            # Nothing found
            return ''

    def description(self, group):
        """Get a description for a single group.  If more than one
        group matches ('group' is a pattern), return the first.  If no
        group matches, return an empty string.

        This elides the response code from the server, since it can
        only be '215' or '285' (for xgtitle) anyway.  If the response
        code is needed, use the 'descriptions' method.

        NOTE: This neither checks for a wildcard in 'group' nor does
        it check whether the group actually exists."""
        return self._getdescriptions(group, False)

    def descriptions(self, group_pattern):
        """Get descriptions for a range of groups."""
        return self._getdescriptions(group_pattern, True)

    def group(self, name):
        """Process a GROUP command.  Argument:
        - group: the group name
        Returns:
        - resp: server response if successful
        - count: number of articles
        - first: first article number
        - last: last article number
        - name: the group name
        """
        resp = self._shortcmd('GROUP ' + name)
        if not resp.startswith('211'):
            raise NNTPReplyError(resp)
        words = resp.split()
        count = first = last = 0
        n = len(words)
        if n > 1:
            count = words[1]
            if n > 2:
                first = words[2]
                if n > 3:
                    last = words[3]
                    if n > 4:
                        name = words[4].lower()
        return resp, int(count), int(first), int(last), name

    def help(self, *, file=None):
        """Process a HELP command. Argument:
        - file: Filename string or file object to store the result in
        Returns:
        - resp: server response if successful
        - list: list of strings returned by the server in response to the
                HELP command
        """
        return self._longcmdstring('HELP', file)

    def _statparse(self, resp):
        """Internal: parse the response line of a STAT, NEXT, LAST,
        ARTICLE, HEAD or BODY command."""
        if not resp.startswith('22'):
            raise NNTPReplyError(resp)
        words = resp.split()
        art_num = int(words[1])
        message_id = words[2]
        return resp, art_num, message_id

    def _statcmd(self, line):
        """Internal: process a STAT, NEXT or LAST command."""
        resp = self._shortcmd(line)
        return self._statparse(resp)

    def stat(self, message_spec=None):
        """Process a STAT command.  Argument:
        - message_spec: article number or message id (if not specified,
          the current article is selected)
        Returns:
        - resp: server response if successful
        - art_num: the article number
        - message_id: the message id
        """
        if message_spec:
            return self._statcmd('STAT {0}'.format(message_spec))
        else:
            return self._statcmd('STAT')

    def next(self):
        """Process a NEXT command.  No arguments.  Return as for STAT."""
        return self._statcmd('NEXT')

    def last(self):
        """Process a LAST command.  No arguments.  Return as for STAT."""
        return self._statcmd('LAST')

    def _artcmd(self, line, file=None):
        """Internal: process a HEAD, BODY or ARTICLE command."""
        resp, lines = self._longcmd(line, file)
        resp, art_num, message_id = self._statparse(resp)
        return resp, ArticleInfo(art_num, message_id, lines)

    def head(self, message_spec=None, *, file=None):
        """Process a HEAD command.  Argument:
        - message_spec: article number or message id
        - file: filename string or file object to store the headers in
        Returns:
        - resp: server response if successful
        - ArticleInfo: (article number, message id, list of header lines)
        """
        if message_spec is not None:
            cmd = 'HEAD {0}'.format(message_spec)
        else:
            cmd = 'HEAD'
        return self._artcmd(cmd, file)

    def body(self, message_spec=None, *, file=None):
        """Process a BODY command.  Argument:
        - message_spec: article number or message id
        - file: filename string or file object to store the body in
        Returns:
        - resp: server response if successful
        - ArticleInfo: (article number, message id, list of body lines)
        """
        if message_spec is not None:
            cmd = 'BODY {0}'.format(message_spec)
        else:
            cmd = 'BODY'
        return self._artcmd(cmd, file)

    def article(self, message_spec=None, *, file=None):
        """Process an ARTICLE command.  Argument:
        - message_spec: article number or message id
        - file: filename string or file object to store the article in
        Returns:
        - resp: server response if successful
        - ArticleInfo: (article number, message id, list of article lines)
        """
        if message_spec is not None:
            cmd = 'ARTICLE {0}'.format(message_spec)
        else:
            cmd = 'ARTICLE'
        return self._artcmd(cmd, file)

    def slave(self):
        """Process a SLAVE command.  Returns:
        - resp: server response if successful
        """
        return self._shortcmd('SLAVE')

    def xhdr(self, hdr, str, *, file=None):
        """Process an XHDR command (optional server extension).  Arguments:
        - hdr: the header type (e.g. 'subject')
        - str: an article nr, a message id, or a range nr1-nr2
        - file: Filename string or file object to store the result in
        Returns:
        - resp: server response if successful
        - list: list of (nr, value) strings
        """
        pat = re.compile('^([0-9]+) ?(.*)\n?')
        resp, lines = self._longcmdstring('XHDR {0} {1}'.format(hdr, str), file)
        def remove_number(line):
            m = pat.match(line)
            return m.group(1, 2) if m else line
        return resp, [remove_number(line) for line in lines]

    def xover(self, start, end, *, file=None):
        """Process an XOVER command (optional server extension) Arguments:
        - start: start of range
        - end: end of range
        - file: Filename string or file object to store the result in
        Returns:
        - resp: server response if successful
        - list: list of dicts containing the response fields
        """
        resp, lines = self._longcmdstring('XOVER {0}-{1}'.format(start, end),
                                          file)
        fmt = self._getoverviewfmt()
        return resp, _parse_overview(lines, fmt)

    def over(self, message_spec, *, file=None):
        """Process an OVER command.  If the command isn't supported, fall
        back to XOVER. Arguments:
        - message_spec:
            - either a message id, indicating the article to fetch
              information about
            - or a (start, end) tuple, indicating a range of article numbers;
              if end is None, information up to the newest message will be
              retrieved
            - or None, indicating the current article number must be used
        - file: Filename string or file object to store the result in
        Returns:
        - resp: server response if successful
        - list: list of dicts containing the response fields

        NOTE: the "message id" form isn't supported by XOVER
        """
        cmd = 'OVER' if 'OVER' in self._caps else 'XOVER'
        if isinstance(message_spec, (tuple, list)):
            start, end = message_spec
            cmd += ' {0}-{1}'.format(start, end or '')
        elif message_spec is not None:
            cmd = cmd + ' ' + message_spec
        resp, lines = self._longcmdstring(cmd, file)
        fmt = self._getoverviewfmt()
        return resp, _parse_overview(lines, fmt)

    def date(self):
        """Process the DATE command.
        Returns:
        - resp: server response if successful
        - date: datetime object
        """
        resp = self._shortcmd("DATE")
        if not resp.startswith('111'):
            raise NNTPReplyError(resp)
        elem = resp.split()
        if len(elem) != 2:
            raise NNTPDataError(resp)
        date = elem[1]
        if len(date) != 14:
            raise NNTPDataError(resp)
        return resp, _parse_datetime(date, None)

    def _post(self, command, f):
        resp = self._shortcmd(command)
        # Raises a specific exception if posting is not allowed
        if not resp.startswith('3'):
            raise NNTPReplyError(resp)
        if isinstance(f, (bytes, bytearray)):
            f = f.splitlines()
        # We don't use _putline() because:
        # - we don't want additional CRLF if the file or iterable is already
        #   in the right format
        # - we don't want a spurious flush() after each line is written
        for line in f:
            if not line.endswith(_CRLF):
                line = line.rstrip(b"\r\n") + _CRLF
            if line.startswith(b'.'):
                line = b'.' + line
            self.file.write(line)
        self.file.write(b".\r\n")
        self.file.flush()
        return self._getresp()

    def post(self, data):
        """Process a POST command.  Arguments:
        - data: bytes object, iterable or file containing the article
        Returns:
        - resp: server response if successful"""
        return self._post('POST', data)

    def ihave(self, message_id, data):
        """Process an IHAVE command.  Arguments:
        - message_id: message-id of the article
        - data: file containing the article
        Returns:
        - resp: server response if successful
        Note that if the server refuses the article an exception is raised."""
        return self._post('IHAVE {0}'.format(message_id), data)

    def _close(self):
        self.file.close()
        del self.file

    def quit(self):
        """Process a QUIT command and close the socket.  Returns:
        - resp: server response if successful"""
        try:
            resp = self._shortcmd('QUIT')
        finally:
            self._close()
        return resp

    def login(self, user=None, password=None, usenetrc=True):
        if self.authenticated:
            raise ValueError("Already logged in.")
        if not user and not usenetrc:
            raise ValueError(
                "At least one of `user` and `usenetrc` must be specified")
        # If no login/password was specified but netrc was requested,
        # try to get them from ~/.netrc
        # Presume that if .netrc has an entry, NNRP authentication is required.
        try:
            if usenetrc and not user:
                import netrc
                credentials = netrc.netrc()
                auth = credentials.authenticators(self.host)
                if auth:
                    user = auth[0]
                    password = auth[2]
        except OSError:
            pass
        # Perform NNTP authentication if needed.
        if not user:
            return
        resp = self._shortcmd('authinfo user ' + user)
        if resp.startswith('381'):
            if not password:
                raise NNTPReplyError(resp)
            else:
                resp = self._shortcmd('authinfo pass ' + password)
                if not resp.startswith('281'):
                    raise NNTPPermanentError(resp)
        # Capabilities might have changed after login
        self._caps = None
        self.getcapabilities()
        # Attempt to send mode reader if it was requested after login.
        # Only do so if we're not in reader mode already.
        if self.readermode_afterauth and 'READER' not in self._caps:
            self._setreadermode()
            # Capabilities might have changed after MODE READER
            self._caps = None
            self.getcapabilities()

    def _setreadermode(self):
        try:
            self.welcome = self._shortcmd('mode reader')
        except NNTPPermanentError:
            # Error 5xx, probably 'not implemented'
            pass
        except NNTPTemporaryError as e:
            if e.response.startswith('480'):
                # Need authorization before 'mode reader'
                self.readermode_afterauth = True
            else:
                raise

    if _have_ssl:
        def starttls(self, context=None):
            """Process a STARTTLS command. Arguments:
            - context: SSL context to use for the encrypted connection
            """
            # Per RFC 4642, STARTTLS MUST NOT be sent after authentication or if
            # a TLS session already exists.
            if self.tls_on:
                raise ValueError("TLS is already enabled.")
            if self.authenticated:
                raise ValueError("TLS cannot be started after authentication.")
            resp = self._shortcmd('STARTTLS')
            if resp.startswith('382'):
                self.file.close()
                self.sock = _encrypt_on(self.sock, context, self.host)
                self.file = self.sock.makefile("rwb")
                self.tls_on = True
                # Capabilities may change after TLS starts up, so ask for them
                # again.
                self._caps = None
                self.getcapabilities()
            else:
                raise NNTPError("TLS failed to start.")


class NNTP(_NNTPBase):

    def __init__(self, host, port=NNTP_PORT, user=None, password=None,
                 readermode=None, usenetrc=False,
                 timeout=_GLOBAL_DEFAULT_TIMEOUT):
        """Initialize an instance.  Arguments:
        - host: hostname to connect to
        - port: port to connect to (default the standard NNTP port)
        - user: username to authenticate with
        - password: password to use with username
        - readermode: if true, send 'mode reader' command after
                      connecting.
        - usenetrc: allow loading username and password from ~/.netrc file
                    if not specified explicitly
        - timeout: timeout (in seconds) used for socket connections

        readermode is sometimes necessary if you are connecting to an
        NNTP server on the local machine and intend to call
        reader-specific commands, such as `group'.  If you get
        unexpected NNTPPermanentErrors, you might need to set
        readermode.
        """
        self.host = host
        self.port = port
        self.sock = self._create_socket(timeout)
        file = None
        try:
            file = self.sock.makefile("rwb")
            super().__init__(file, host, readermode, timeout)
            if user or usenetrc:
                self.login(user, password, usenetrc)
        except:
            if file:
                file.close()
            self.sock.close()
            raise

    def _create_socket(self, timeout):
        if timeout is not None and not timeout:
            raise ValueError('Non-blocking socket (timeout=0) is not supported')
        sys.audit("nntplib.connect", self, self.host, self.port)
        return socket.create_connection((self.host, self.port), timeout)

    def _close(self):
        try:
            super()._close()
        finally:
            self.sock.close()


if _have_ssl:
    class NNTP_SSL(NNTP):

        def __init__(self, host, port=NNTP_SSL_PORT,
                    user=None, password=None, ssl_context=None,
                    readermode=None, usenetrc=False,
                    timeout=_GLOBAL_DEFAULT_TIMEOUT):
            """This works identically to NNTP.__init__, except for the change
            in default port and the `ssl_context` argument for SSL connections.
            """
            self.ssl_context = ssl_context
            super().__init__(host, port, user, password, readermode,
                             usenetrc, timeout)

        def _create_socket(self, timeout):
            sock = super()._create_socket(timeout)
            try:
                sock = _encrypt_on(sock, self.ssl_context, self.host)
            except:
                sock.close()
                raise
            else:
                return sock

    __all__.append("NNTP_SSL")


# Test retrieval when run as a script.
if __name__ == '__main__':
    import argparse

    parser = argparse.ArgumentParser(description="""\
        nntplib built-in demo - display the latest articles in a newsgroup""")
    parser.add_argument('-g', '--group', default='gmane.comp.python.general',
                        help='group to fetch messages from (default: %(default)s)')
    parser.add_argument('-s', '--server', default='news.gmane.io',
                        help='NNTP server hostname (default: %(default)s)')
    parser.add_argument('-p', '--port', default=-1, type=int,
                        help='NNTP port number (default: %s / %s)' % (NNTP_PORT, NNTP_SSL_PORT))
    parser.add_argument('-n', '--nb-articles', default=10, type=int,
                        help='number of articles to fetch (default: %(default)s)')
    parser.add_argument('-S', '--ssl', action='store_true', default=False,
                        help='use NNTP over SSL')
    args = parser.parse_args()

    port = args.port
    if not args.ssl:
        if port == -1:
            port = NNTP_PORT
        s = NNTP(host=args.server, port=port)
    else:
        if port == -1:
            port = NNTP_SSL_PORT
        s = NNTP_SSL(host=args.server, port=port)

    caps = s.getcapabilities()
    if 'STARTTLS' in caps:
        s.starttls()
    resp, count, first, last, name = s.group(args.group)
    print('Group', name, 'has', count, 'articles, range', first, 'to', last)

    def cut(s, lim):
        if len(s) > lim:
            s = s[:lim - 4] + "..."
        return s

    first = str(int(last) - args.nb_articles + 1)
    resp, overviews = s.xover(first, last)
    for artnum, over in overviews:
        author = decode_header(over['from']).split('<', 1)[0]
        subject = decode_header(over['subject'])
        lines = int(over[':lines'])
        print("{:7} {:20} {:42} ({})".format(
              artnum, cut(author, 20), cut(subject, 42), lines)
              )

    s.quit()