compat.py

# -*- test-case-name: twisted.test.test_compat -*-
#
# Copyright (c) Twisted Matrix Laboratories.
# See LICENSE for details.


"""
Compatibility module to provide backwards compatibility for useful Python
features.

This is mainly for use of internal Twisted code. We encourage you to use
the latest version of Python directly from your code, if possible.

@var unicode: The type of Unicode strings, C{unicode} on Python 2 and C{str}
    on Python 3.

@var NativeStringIO: An in-memory file-like object that operates on the native
    string type (bytes in Python 2, unicode in Python 3).
"""

from __future__ import division

import sys, string, socket, struct


if sys.version_info < (3, 0):
    _PY3 = False
else:
    _PY3 = True



def inet_pton(af, addr):
    if af == socket.AF_INET:
        return socket.inet_aton(addr)
    elif af == getattr(socket, 'AF_INET6', 'AF_INET6'):
        if [x for x in addr if x not in string.hexdigits + ':.']:
            raise ValueError("Illegal characters: %r" % (''.join(x),))

        parts = addr.split(':')
        elided = parts.count('')
        ipv4Component = '.' in parts[-1]

        if len(parts) > (8 - ipv4Component) or elided > 3:
            raise ValueError("Syntactically invalid address")

        if elided == 3:
            return '\x00' * 16

        if elided:
            zeros = ['0'] * (8 - len(parts) - ipv4Component + elided)

            if addr.startswith('::'):
                parts[:2] = zeros
            elif addr.endswith('::'):
                parts[-2:] = zeros
            else:
                idx = parts.index('')
                parts[idx:idx+1] = zeros

            if len(parts) != 8 - ipv4Component:
                raise ValueError("Syntactically invalid address")
        else:
            if len(parts) != (8 - ipv4Component):
                raise ValueError("Syntactically invalid address")

        if ipv4Component:
            if parts[-1].count('.') != 3:
                raise ValueError("Syntactically invalid address")
            rawipv4 = socket.inet_aton(parts[-1])
            unpackedipv4 = struct.unpack('!HH', rawipv4)
            parts[-1:] = [hex(x)[2:] for x in unpackedipv4]

        parts = [int(x, 16) for x in parts]
        return struct.pack('!8H', *parts)
    else:
        raise socket.error(97, 'Address family not supported by protocol')

def inet_ntop(af, addr):
    if af == socket.AF_INET:
        return socket.inet_ntoa(addr)
    elif af == socket.AF_INET6:
        if len(addr) != 16:
            raise ValueError("address length incorrect")
        parts = struct.unpack('!8H', addr)
        curBase = bestBase = None
        for i in range(8):
            if not parts[i]:
                if curBase is None:
                    curBase = i
                    curLen = 0
                curLen += 1
            else:
                if curBase is not None:
                    if bestBase is None or curLen > bestLen:
                        bestBase = curBase
                        bestLen = curLen
                    curBase = None
        if curBase is not None and (bestBase is None or curLen > bestLen):
            bestBase = curBase
            bestLen = curLen
        parts = [hex(x)[2:] for x in parts]
        if bestBase is not None:
            parts[bestBase:bestBase + bestLen] = ['']
        if parts[0] == '':
            parts.insert(0, '')
        if parts[-1] == '':
            parts.insert(len(parts) - 1, '')
        return ':'.join(parts)
    else:
        raise socket.error(97, 'Address family not supported by protocol')

try:
    socket.AF_INET6
except AttributeError:
    socket.AF_INET6 = 'AF_INET6'

try:
    socket.inet_pton(socket.AF_INET6, "::")
except (AttributeError, NameError, socket.error):
    socket.inet_pton = inet_pton
    socket.inet_ntop = inet_ntop


adict = dict



if _PY3:
    # These are actually useless in Python 2 as well, but we need to go
    # through deprecation process there (ticket #5895):
    del adict, inet_pton, inet_ntop



try:
    set = set
except NameError:
    from sets import Set as set


try:
    frozenset = frozenset
except NameError:
    from sets import ImmutableSet as frozenset


try:
    from functools import reduce
except ImportError:
    reduce = reduce



def execfile(filename, globals, locals=None):
    """
    Execute a Python script in the given namespaces.

    Similar to the execfile builtin, but a namespace is mandatory, partly
    because that's a sensible thing to require, and because otherwise we'd
    have to do some frame hacking.

    This is a compatibility implementation for Python 3 porting, to avoid the
    use of the deprecated builtin C{execfile} function.
    """
    if locals is None:
        locals = globals
    fin = open(filename, "rbU")
    try:
        source = fin.read()
    finally:
        fin.close()
    code = compile(source, filename, "exec")
    exec(code, globals, locals)


try:
    cmp = cmp
except NameError:
    def cmp(a, b):
        """
        Compare two objects.

        Returns a negative number if C{a < b}, zero if they are equal, and a
        positive number if C{a > b}.
        """
        if a < b:
            return -1
        elif a == b:
            return 0
        else:
            return 1



def comparable(klass):
    """
    Class decorator that ensures support for the special C{__cmp__} method.

    On Python 2 this does nothing.

    On Python 3, C{__eq__}, C{__lt__}, etc. methods are added to the class,
    relying on C{__cmp__} to implement their comparisons.
    """
    # On Python 2, __cmp__ will just work, so no need to add extra methods:
    if not _PY3:
        return klass

    def __eq__(self, other):
        c = self.__cmp__(other)
        if c is NotImplemented:
            return c
        return c == 0


    def __ne__(self, other):
        c = self.__cmp__(other)
        if c is NotImplemented:
            return c
        return c != 0


    def __lt__(self, other):
        c = self.__cmp__(other)
        if c is NotImplemented:
            return c
        return c < 0


    def __le__(self, other):
        c = self.__cmp__(other)
        if c is NotImplemented:
            return c
        return c <= 0


    def __gt__(self, other):
        c = self.__cmp__(other)
        if c is NotImplemented:
            return c
        return c > 0


    def __ge__(self, other):
        c = self.__cmp__(other)
        if c is NotImplemented:
            return c
        return c >= 0

    klass.__lt__ = __lt__
    klass.__gt__ = __gt__
    klass.__le__ = __le__
    klass.__ge__ = __ge__
    klass.__eq__ = __eq__
    klass.__ne__ = __ne__
    return klass



if _PY3:
    unicode = str
else:
    unicode = unicode



def nativeString(s):
    """
    Convert C{bytes} or C{unicode} to the native C{str} type, using ASCII
    encoding if conversion is necessary.

    @raise UnicodeError: The input string is not ASCII encodable/decodable.
    @raise TypeError: The input is neither C{bytes} nor C{unicode}.
    """
    if not isinstance(s, (bytes, unicode)):
        raise TypeError("%r is neither bytes nor unicode" % s)
    if _PY3:
        if isinstance(s, bytes):
            return s.decode("ascii")
        else:
            # Ensure we're limited to ASCII subset:
            s.encode("ascii")
    else:
        if isinstance(s, unicode):
            return s.encode("ascii")
        else:
            # Ensure we're limited to ASCII subset:
            s.decode("ascii")
    return s



if _PY3:
    def reraise(exception, traceback):
        raise exception.with_traceback(traceback)
else:
    exec("""def reraise(exception, traceback):
        raise exception.__class__, exception, traceback""")

reraise.__doc__ = """
Re-raise an exception, with an optional traceback, in a way that is compatible
with both Python 2 and Python 3.

Note that on Python 3, re-raised exceptions will be mutated, with their
C{__traceback__} attribute being set.

@param exception: The exception instance.
@param traceback: The traceback to use, or C{None} indicating a new traceback.
"""



if _PY3:
    from io import StringIO as NativeStringIO
else:
    from io import BytesIO as NativeStringIO



# Functions for dealing with Python 3's bytes type, which is somewhat
# different than Python 2's:
if _PY3:
    def iterbytes(originalBytes):
        for i in range(len(originalBytes)):
            yield originalBytes[i:i+1]


    def intToBytes(i):
        return ("%d" % i).encode("ascii")


    # Ideally we would use memoryview, but it has a number of differences from
    # the Python 2 buffer() that make that impractical
    # (http://bugs.python.org/issue15945, incompatiblity with pyOpenSSL due to
    # PyArg_ParseTuple differences.)
    def lazyByteSlice(object, offset=0, size=None):
        """
        Return a copy of the given bytes-like object.

        If an offset is given, the copy starts at that offset. If a size is
        given, the copy will only be of that length.

        @param object: C{bytes} to be copied.

        @param offset: C{int}, starting index of copy.

        @param size: Optional, if an C{int} is given limit the length of copy
            to this size.
        """
        if size is None:
            return object[offset:]
        else:
            return object[offset:(offset + size)]


    def networkString(s):
        if not isinstance(s, unicode):
            raise TypeError("Can only convert text to bytes on Python 3")
        return s.encode('ascii')
else:
    def iterbytes(originalBytes):
        return originalBytes


    def intToBytes(i):
        return b"%d" % i


    lazyByteSlice = buffer

    def networkString(s):
        if not isinstance(s, str):
            raise TypeError("Can only pass-through bytes on Python 2")
        # Ensure we're limited to ASCII subset:
        s.decode('ascii')
        return s

iterbytes.__doc__ = """
Return an iterable wrapper for a C{bytes} object that provides the behavior of
iterating over C{bytes} on Python 2.

In particular, the results of iteration are the individual bytes (rather than
integers as on Python 3).

@param originalBytes: A C{bytes} object that will be wrapped.
"""

intToBytes.__doc__ = """
Convert the given integer into C{bytes}, as ASCII-encoded Arab numeral.

In other words, this is equivalent to calling C{bytes} in Python 2 on an
integer.

@param i: The C{int} to convert to C{bytes}.
@rtype: C{bytes}
"""

networkString.__doc__ = """
Convert the native string type to C{bytes} if it is not already C{bytes} using
ASCII encoding if conversion is necessary.

This is useful for sending text-like bytes that are constructed using string
interpolation.  For example, this is safe on Python 2 and Python 3:

    networkString("Hello %d" % (n,))

@param s: A native string to convert to bytes if necessary.
@type s: C{str}

@raise UnicodeError: The input string is not ASCII encodable/decodable.
@raise TypeError: The input is neither C{bytes} nor C{unicode}.

@rtype: C{bytes}
"""


__all__ = [
    "reraise",
    "execfile",
    "frozenset",
    "reduce",
    "set",
    "cmp",
    "comparable",
    "nativeString",
    "NativeStringIO",
    "networkString",
    "unicode",
    "iterbytes",
    "intToBytes",
    "lazyByteSlice",
    ]