From c5cf7973422dce0ed59849aaf2d708d4c6b7320d Mon Sep 17 00:00:00 2001 From: Ronald Oussoren Date: Thu, 21 Nov 2013 15:46:49 +0100 Subject: Issue #14455: plistlib now supports binary plists and has an updated API. This patch adds support for binary plists on OSX to plistlib (based on a patch by 'dpounces'). The patch also cleans up the API for the plistlib module. --- Lib/plistlib.py | 1098 +++++++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 827 insertions(+), 271 deletions(-) (limited to 'Lib/plistlib.py') diff --git a/Lib/plistlib.py b/Lib/plistlib.py index 2b0b634100..bdb8ab352e 100644 --- a/Lib/plistlib.py +++ b/Lib/plistlib.py @@ -4,25 +4,20 @@ The property list (.plist) file format is a simple XML pickle supporting basic object types, like dictionaries, lists, numbers and strings. Usually the top level object is a dictionary. -To write out a plist file, use the writePlist(rootObject, pathOrFile) -function. 'rootObject' is the top level object, 'pathOrFile' is a -filename or a (writable) file object. +To write out a plist file, use the dump(value, file) +function. 'value' is the top level object, 'file' is +a (writable) file object. -To parse a plist from a file, use the readPlist(pathOrFile) function, -with a file name or a (readable) file object as the only argument. It +To parse a plist from a file, use the load(file) function, +with a (readable) file object as the only argument. It returns the top level object (again, usually a dictionary). -To work with plist data in bytes objects, you can use readPlistFromBytes() -and writePlistToBytes(). +To work with plist data in bytes objects, you can use loads() +and dumps(). Values can be strings, integers, floats, booleans, tuples, lists, -dictionaries (but only with string keys), Data or datetime.datetime objects. -String values (including dictionary keys) have to be unicode strings -- they -will be written out as UTF-8. - -The plist type is supported through the Data class. This is a -thin wrapper around a Python bytes object. Use 'Data' if your strings -contain control characters. +dictionaries (but only with string keys), Data, bytes, bytearray, or +datetime.datetime objects. Generate Plist example: @@ -37,226 +32,48 @@ Generate Plist example: aTrueValue = True, aFalseValue = False, ), - someData = Data(b""), - someMoreData = Data(b"" * 10), + someData = b"", + someMoreData = b"" * 10, aDate = datetime.datetime.fromtimestamp(time.mktime(time.gmtime())), ) - writePlist(pl, fileName) + with open(fileName, 'wb') as fp: + dump(pl, fp) Parse Plist example: - pl = readPlist(pathOrFile) - print pl["aKey"] + with open(fileName, 'rb') as fp: + pl = load(fp) + print(pl["aKey"]) """ - - __all__ = [ "readPlist", "writePlist", "readPlistFromBytes", "writePlistToBytes", - "Plist", "Data", "Dict" + "Plist", "Data", "Dict", "FMT_XML", "FMT_BINARY", + "load", "dump", "loads", "dumps" ] -# Note: the Plist and Dict classes have been deprecated. import binascii +import codecs +import contextlib import datetime +import enum from io import BytesIO +import itertools +import os import re +import struct +from warnings import warn +from xml.parsers.expat import ParserCreate -def readPlist(pathOrFile): - """Read a .plist file. 'pathOrFile' may either be a file name or a - (readable) file object. Return the unpacked root object (which - usually is a dictionary). - """ - didOpen = False - try: - if isinstance(pathOrFile, str): - pathOrFile = open(pathOrFile, 'rb') - didOpen = True - p = PlistParser() - rootObject = p.parse(pathOrFile) - finally: - if didOpen: - pathOrFile.close() - return rootObject - - -def writePlist(rootObject, pathOrFile): - """Write 'rootObject' to a .plist file. 'pathOrFile' may either be a - file name or a (writable) file object. - """ - didOpen = False - try: - if isinstance(pathOrFile, str): - pathOrFile = open(pathOrFile, 'wb') - didOpen = True - writer = PlistWriter(pathOrFile) - writer.writeln("") - writer.writeValue(rootObject) - writer.writeln("") - finally: - if didOpen: - pathOrFile.close() - - -def readPlistFromBytes(data): - """Read a plist data from a bytes object. Return the root object. - """ - return readPlist(BytesIO(data)) - - -def writePlistToBytes(rootObject): - """Return 'rootObject' as a plist-formatted bytes object. - """ - f = BytesIO() - writePlist(rootObject, f) - return f.getvalue() - - -class DumbXMLWriter: - def __init__(self, file, indentLevel=0, indent="\t"): - self.file = file - self.stack = [] - self.indentLevel = indentLevel - self.indent = indent - - def beginElement(self, element): - self.stack.append(element) - self.writeln("<%s>" % element) - self.indentLevel += 1 - - def endElement(self, element): - assert self.indentLevel > 0 - assert self.stack.pop() == element - self.indentLevel -= 1 - self.writeln("" % element) - - def simpleElement(self, element, value=None): - if value is not None: - value = _escape(value) - self.writeln("<%s>%s" % (element, value, element)) - else: - self.writeln("<%s/>" % element) - - def writeln(self, line): - if line: - # plist has fixed encoding of utf-8 - if isinstance(line, str): - line = line.encode('utf-8') - self.file.write(self.indentLevel * self.indent) - self.file.write(line) - self.file.write(b'\n') - - -# Contents should conform to a subset of ISO 8601 -# (in particular, YYYY '-' MM '-' DD 'T' HH ':' MM ':' SS 'Z'. Smaller units may be omitted with -# a loss of precision) -_dateParser = re.compile(r"(?P\d\d\d\d)(?:-(?P\d\d)(?:-(?P\d\d)(?:T(?P\d\d)(?::(?P\d\d)(?::(?P\d\d))?)?)?)?)?Z", re.ASCII) - -def _dateFromString(s): - order = ('year', 'month', 'day', 'hour', 'minute', 'second') - gd = _dateParser.match(s).groupdict() - lst = [] - for key in order: - val = gd[key] - if val is None: - break - lst.append(int(val)) - return datetime.datetime(*lst) - -def _dateToString(d): - return '%04d-%02d-%02dT%02d:%02d:%02dZ' % ( - d.year, d.month, d.day, - d.hour, d.minute, d.second - ) - - -# Regex to find any control chars, except for \t \n and \r -_controlCharPat = re.compile( - r"[\x00\x01\x02\x03\x04\x05\x06\x07\x08\x0b\x0c\x0e\x0f" - r"\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f]") - -def _escape(text): - m = _controlCharPat.search(text) - if m is not None: - raise ValueError("strings can't contains control characters; " - "use plistlib.Data instead") - text = text.replace("\r\n", "\n") # convert DOS line endings - text = text.replace("\r", "\n") # convert Mac line endings - text = text.replace("&", "&") # escape '&' - text = text.replace("<", "<") # escape '<' - text = text.replace(">", ">") # escape '>' - return text - - -PLISTHEADER = b"""\ - - -""" - -class PlistWriter(DumbXMLWriter): - - def __init__(self, file, indentLevel=0, indent=b"\t", writeHeader=1): - if writeHeader: - file.write(PLISTHEADER) - DumbXMLWriter.__init__(self, file, indentLevel, indent) - - def writeValue(self, value): - if isinstance(value, str): - self.simpleElement("string", value) - elif isinstance(value, bool): - # must switch for bool before int, as bool is a - # subclass of int... - if value: - self.simpleElement("true") - else: - self.simpleElement("false") - elif isinstance(value, int): - self.simpleElement("integer", "%d" % value) - elif isinstance(value, float): - self.simpleElement("real", repr(value)) - elif isinstance(value, dict): - self.writeDict(value) - elif isinstance(value, Data): - self.writeData(value) - elif isinstance(value, datetime.datetime): - self.simpleElement("date", _dateToString(value)) - elif isinstance(value, (tuple, list)): - self.writeArray(value) - else: - raise TypeError("unsupported type: %s" % type(value)) - - def writeData(self, data): - self.beginElement("data") - self.indentLevel -= 1 - maxlinelength = max(16, 76 - len(self.indent.replace(b"\t", b" " * 8) * - self.indentLevel)) - for line in data.asBase64(maxlinelength).split(b"\n"): - if line: - self.writeln(line) - self.indentLevel += 1 - self.endElement("data") +PlistFormat = enum.Enum('PlistFormat', 'FMT_XML FMT_BINARY', module=__name__) +globals().update(PlistFormat.__members__) - def writeDict(self, d): - if d: - self.beginElement("dict") - items = sorted(d.items()) - for key, value in items: - if not isinstance(key, str): - raise TypeError("keys must be strings") - self.simpleElement("key", key) - self.writeValue(value) - self.endElement("dict") - else: - self.simpleElement("dict") - def writeArray(self, array): - if array: - self.beginElement("array") - for value in array: - self.writeValue(value) - self.endElement("array") - else: - self.simpleElement("array") +# +# +# Deprecated functionality +# +# class _InternalDict(dict): @@ -264,19 +81,18 @@ class _InternalDict(dict): # This class is needed while Dict is scheduled for deprecation: # we only need to warn when a *user* instantiates Dict or when # the "attribute notation for dict keys" is used. + __slots__ = () def __getattr__(self, attr): try: value = self[attr] except KeyError: raise AttributeError(attr) - from warnings import warn warn("Attribute access from plist dicts is deprecated, use d[key] " "notation instead", DeprecationWarning, 2) return value def __setattr__(self, attr, value): - from warnings import warn warn("Attribute access from plist dicts is deprecated, use d[key] " "notation instead", DeprecationWarning, 2) self[attr] = value @@ -286,56 +102,111 @@ class _InternalDict(dict): del self[attr] except KeyError: raise AttributeError(attr) - from warnings import warn warn("Attribute access from plist dicts is deprecated, use d[key] " "notation instead", DeprecationWarning, 2) + class Dict(_InternalDict): def __init__(self, **kwargs): - from warnings import warn warn("The plistlib.Dict class is deprecated, use builtin dict instead", DeprecationWarning, 2) super().__init__(**kwargs) -class Plist(_InternalDict): +@contextlib.contextmanager +def _maybe_open(pathOrFile, mode): + if isinstance(pathOrFile, str): + with open(pathOrFile, mode) as fp: + yield fp + + else: + yield pathOrFile + - """This class has been deprecated. Use readPlist() and writePlist() +class Plist(_InternalDict): + """This class has been deprecated. Use dump() and load() functions instead, together with regular dict objects. """ def __init__(self, **kwargs): - from warnings import warn - warn("The Plist class is deprecated, use the readPlist() and " - "writePlist() functions instead", DeprecationWarning, 2) + warn("The Plist class is deprecated, use the load() and " + "dump() functions instead", DeprecationWarning, 2) super().__init__(**kwargs) + @classmethod def fromFile(cls, pathOrFile): - """Deprecated. Use the readPlist() function instead.""" - rootObject = readPlist(pathOrFile) + """Deprecated. Use the load() function instead.""" + with maybe_open(pathOrFile, 'rb') as fp: + value = load(fp) plist = cls() - plist.update(rootObject) + plist.update(value) return plist - fromFile = classmethod(fromFile) def write(self, pathOrFile): - """Deprecated. Use the writePlist() function instead.""" - writePlist(self, pathOrFile) + """Deprecated. Use the dump() function instead.""" + with _maybe_open(pathOrFile, 'wb') as fp: + dump(self, fp) -def _encodeBase64(s, maxlinelength=76): - # copied from base64.encodebytes(), with added maxlinelength argument - maxbinsize = (maxlinelength//4)*3 - pieces = [] - for i in range(0, len(s), maxbinsize): - chunk = s[i : i + maxbinsize] - pieces.append(binascii.b2a_base64(chunk)) - return b''.join(pieces) +def readPlist(pathOrFile): + """ + Read a .plist from a path or file. pathOrFile should either + be a file name, or a readable binary file object. + + This function is deprecated, use load instead. + """ + warn("The readPlist function is deprecated, use load() instead", + DeprecationWarning, 2) + + with _maybe_open(pathOrFile, 'rb') as fp: + return load(fp, fmt=None, use_builtin_types=False, + dict_type=_InternalDict) + +def writePlist(value, pathOrFile): + """ + Write 'value' to a .plist file. 'pathOrFile' may either be a + file name or a (writable) file object. + + This function is deprecated, use dump instead. + """ + warn("The writePlist function is deprecated, use dump() instead", + DeprecationWarning, 2) + with _maybe_open(pathOrFile, 'wb') as fp: + dump(value, fp, fmt=FMT_XML, sort_keys=True, skipkeys=False) + + +def readPlistFromBytes(data): + """ + Read a plist data from a bytes object. Return the root object. + + This function is deprecated, use loads instead. + """ + warn("The readPlistFromBytes function is deprecated, use loads() instead", + DeprecationWarning, 2) + return load(BytesIO(data), fmt=None, use_builtin_types=False, + dict_type=_InternalDict) + + +def writePlistToBytes(value): + """ + Return 'value' as a plist-formatted bytes object. + + This function is deprecated, use dumps instead. + """ + warn("The writePlistToBytes function is deprecated, use dumps() instead", + DeprecationWarning, 2) + f = BytesIO() + dump(value, f, fmt=FMT_XML, sort_keys=True, skipkeys=False) + return f.getvalue() + class Data: + """ + Wrapper for binary data. - """Wrapper for binary data.""" + This class is deprecated, use a bytes object instead. + """ def __init__(self, data): if not isinstance(data, bytes): @@ -346,10 +217,10 @@ class Data: def fromBase64(cls, data): # base64.decodebytes just calls binascii.a2b_base64; # it seems overkill to use both base64 and binascii. - return cls(binascii.a2b_base64(data)) + return cls(_decode_base64(data)) def asBase64(self, maxlinelength=76): - return _encodeBase64(self.data, maxlinelength) + return _encode_base64(self.data, maxlinelength) def __eq__(self, other): if isinstance(other, self.__class__): @@ -362,43 +233,119 @@ class Data: def __repr__(self): return "%s(%s)" % (self.__class__.__name__, repr(self.data)) -class PlistParser: +# +# +# End of deprecated functionality +# +# + + +# +# XML support +# + + +# XML 'header' +PLISTHEADER = b"""\ + + +""" + + +# Regex to find any control chars, except for \t \n and \r +_controlCharPat = re.compile( + r"[\x00\x01\x02\x03\x04\x05\x06\x07\x08\x0b\x0c\x0e\x0f" + r"\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f]") + +def _encode_base64(s, maxlinelength=76): + # copied from base64.encodebytes(), with added maxlinelength argument + maxbinsize = (maxlinelength//4)*3 + pieces = [] + for i in range(0, len(s), maxbinsize): + chunk = s[i : i + maxbinsize] + pieces.append(binascii.b2a_base64(chunk)) + return b''.join(pieces) + +def _decode_base64(s): + if isinstance(s, str): + return binascii.a2b_base64(s.encode("utf-8")) + + else: + return binascii.a2b_base64(s) - def __init__(self): +# Contents should conform to a subset of ISO 8601 +# (in particular, YYYY '-' MM '-' DD 'T' HH ':' MM ':' SS 'Z'. Smaller units +# may be omitted with # a loss of precision) +_dateParser = re.compile(r"(?P\d\d\d\d)(?:-(?P\d\d)(?:-(?P\d\d)(?:T(?P\d\d)(?::(?P\d\d)(?::(?P\d\d))?)?)?)?)?Z", re.ASCII) + + +def _date_from_string(s): + order = ('year', 'month', 'day', 'hour', 'minute', 'second') + gd = _dateParser.match(s).groupdict() + lst = [] + for key in order: + val = gd[key] + if val is None: + break + lst.append(int(val)) + return datetime.datetime(*lst) + + +def _date_to_string(d): + return '%04d-%02d-%02dT%02d:%02d:%02dZ' % ( + d.year, d.month, d.day, + d.hour, d.minute, d.second + ) + +def _escape(text): + m = _controlCharPat.search(text) + if m is not None: + raise ValueError("strings can't contains control characters; " + "use bytes instead") + text = text.replace("\r\n", "\n") # convert DOS line endings + text = text.replace("\r", "\n") # convert Mac line endings + text = text.replace("&", "&") # escape '&' + text = text.replace("<", "<") # escape '<' + text = text.replace(">", ">") # escape '>' + return text + +class _PlistParser: + def __init__(self, use_builtin_types, dict_type): self.stack = [] - self.currentKey = None + self.current_key = None self.root = None + self._use_builtin_types = use_builtin_types + self._dict_type = dict_type def parse(self, fileobj): - from xml.parsers.expat import ParserCreate self.parser = ParserCreate() - self.parser.StartElementHandler = self.handleBeginElement - self.parser.EndElementHandler = self.handleEndElement - self.parser.CharacterDataHandler = self.handleData + self.parser.StartElementHandler = self.handle_begin_element + self.parser.EndElementHandler = self.handle_end_element + self.parser.CharacterDataHandler = self.handle_data self.parser.ParseFile(fileobj) return self.root - def handleBeginElement(self, element, attrs): + def handle_begin_element(self, element, attrs): self.data = [] handler = getattr(self, "begin_" + element, None) if handler is not None: handler(attrs) - def handleEndElement(self, element): + def handle_end_element(self, element): handler = getattr(self, "end_" + element, None) if handler is not None: handler() - def handleData(self, data): + def handle_data(self, data): self.data.append(data) - def addObject(self, value): - if self.currentKey is not None: + def add_object(self, value): + if self.current_key is not None: if not isinstance(self.stack[-1], type({})): raise ValueError("unexpected element at line %d" % self.parser.CurrentLineNumber) - self.stack[-1][self.currentKey] = value - self.currentKey = None + self.stack[-1][self.current_key] = value + self.current_key = None elif not self.stack: # this is the root object self.root = value @@ -408,7 +355,7 @@ class PlistParser: self.parser.CurrentLineNumber) self.stack[-1].append(value) - def getData(self): + def get_data(self): data = ''.join(self.data) self.data = [] return data @@ -416,39 +363,648 @@ class PlistParser: # element handlers def begin_dict(self, attrs): - d = _InternalDict() - self.addObject(d) + d = self._dict_type() + self.add_object(d) self.stack.append(d) + def end_dict(self): - if self.currentKey: + if self.current_key: raise ValueError("missing value for key '%s' at line %d" % - (self.currentKey,self.parser.CurrentLineNumber)) + (self.current_key,self.parser.CurrentLineNumber)) self.stack.pop() def end_key(self): - if self.currentKey or not isinstance(self.stack[-1], type({})): + if self.current_key or not isinstance(self.stack[-1], type({})): raise ValueError("unexpected key at line %d" % self.parser.CurrentLineNumber) - self.currentKey = self.getData() + self.current_key = self.get_data() def begin_array(self, attrs): a = [] - self.addObject(a) + self.add_object(a) self.stack.append(a) + def end_array(self): self.stack.pop() def end_true(self): - self.addObject(True) + self.add_object(True) + def end_false(self): - self.addObject(False) + self.add_object(False) + def end_integer(self): - self.addObject(int(self.getData())) + self.add_object(int(self.get_data())) + def end_real(self): - self.addObject(float(self.getData())) + self.add_object(float(self.get_data())) + def end_string(self): - self.addObject(self.getData()) + self.add_object(self.get_data()) + def end_data(self): - self.addObject(Data.fromBase64(self.getData().encode("utf-8"))) + if self._use_builtin_types: + self.add_object(_decode_base64(self.get_data())) + + else: + self.add_object(Data.fromBase64(self.get_data())) + def end_date(self): - self.addObject(_dateFromString(self.getData())) + self.add_object(_date_from_string(self.get_data())) + + +class _DumbXMLWriter: + def __init__(self, file, indent_level=0, indent="\t"): + self.file = file + self.stack = [] + self._indent_level = indent_level + self.indent = indent + + def begin_element(self, element): + self.stack.append(element) + self.writeln("<%s>" % element) + self._indent_level += 1 + + def end_element(self, element): + assert self._indent_level > 0 + assert self.stack.pop() == element + self._indent_level -= 1 + self.writeln("" % element) + + def simple_element(self, element, value=None): + if value is not None: + value = _escape(value) + self.writeln("<%s>%s" % (element, value, element)) + + else: + self.writeln("<%s/>" % element) + + def writeln(self, line): + if line: + # plist has fixed encoding of utf-8 + + # XXX: is this test needed? + if isinstance(line, str): + line = line.encode('utf-8') + self.file.write(self._indent_level * self.indent) + self.file.write(line) + self.file.write(b'\n') + + +class _PlistWriter(_DumbXMLWriter): + def __init__( + self, file, indent_level=0, indent=b"\t", writeHeader=1, + sort_keys=True, skipkeys=False): + + if writeHeader: + file.write(PLISTHEADER) + _DumbXMLWriter.__init__(self, file, indent_level, indent) + self._sort_keys = sort_keys + self._skipkeys = skipkeys + + def write(self, value): + self.writeln("") + self.write_value(value) + self.writeln("") + + def write_value(self, value): + if isinstance(value, str): + self.simple_element("string", value) + + elif value is True: + self.simple_element("true") + + elif value is False: + self.simple_element("false") + + elif isinstance(value, int): + self.simple_element("integer", "%d" % value) + + elif isinstance(value, float): + self.simple_element("real", repr(value)) + + elif isinstance(value, dict): + self.write_dict(value) + + elif isinstance(value, Data): + self.write_data(value) + + elif isinstance(value, (bytes, bytearray)): + self.write_bytes(value) + + elif isinstance(value, datetime.datetime): + self.simple_element("date", _date_to_string(value)) + + elif isinstance(value, (tuple, list)): + self.write_array(value) + + else: + raise TypeError("unsupported type: %s" % type(value)) + + def write_data(self, data): + self.write_bytes(data.data) + + def write_bytes(self, data): + self.begin_element("data") + self._indent_level -= 1 + maxlinelength = max( + 16, + 76 - len(self.indent.replace(b"\t", b" " * 8) * self._indent_level)) + + for line in _encode_base64(data, maxlinelength).split(b"\n"): + if line: + self.writeln(line) + self._indent_level += 1 + self.end_element("data") + + def write_dict(self, d): + if d: + self.begin_element("dict") + if self._sort_keys: + items = sorted(d.items()) + else: + items = d.items() + + for key, value in items: + if not isinstance(key, str): + if self._skipkeys: + continue + raise TypeError("keys must be strings") + self.simple_element("key", key) + self.write_value(value) + self.end_element("dict") + + else: + self.simple_element("dict") + + def write_array(self, array): + if array: + self.begin_element("array") + for value in array: + self.write_value(value) + self.end_element("array") + + else: + self.simple_element("array") + + +def _is_fmt_xml(header): + prefixes = (b'offset... + # TRAILER + self._fp = fp + self._fp.seek(-32, os.SEEK_END) + trailer = self._fp.read(32) + if len(trailer) != 32: + raise InvalidFileException() + ( + offset_size, self._ref_size, num_objects, top_object, + offset_table_offset + ) = struct.unpack('>6xBBQQQ', trailer) + self._fp.seek(offset_table_offset) + offset_format = '>' + _BINARY_FORMAT[offset_size] * num_objects + self._ref_format = _BINARY_FORMAT[self._ref_size] + self._object_offsets = struct.unpack( + offset_format, self._fp.read(offset_size * num_objects)) + return self._read_object(self._object_offsets[top_object]) + + except (OSError, IndexError, struct.error): + raise InvalidFileException() + + def _get_size(self, tokenL): + """ return the size of the next object.""" + if tokenL == 0xF: + m = self._fp.read(1)[0] & 0x3 + s = 1 << m + f = '>' + _BINARY_FORMAT[s] + return struct.unpack(f, self._fp.read(s))[0] + + return tokenL + + def _read_refs(self, n): + return struct.unpack( + '>' + self._ref_format * n, self._fp.read(n * self._ref_size)) + + def _read_object(self, offset): + """ + read the object at offset. + + May recursively read sub-objects (content of an array/dict/set) + """ + self._fp.seek(offset) + token = self._fp.read(1)[0] + tokenH, tokenL = token & 0xF0, token & 0x0F + + if token == 0x00: + return None + + elif token == 0x08: + return False + + elif token == 0x09: + return True + + # The referenced source code also mentions URL (0x0c, 0x0d) and + # UUID (0x0e), but neither can be generated using the Cocoa libraries. + + elif token == 0x0f: + return b'' + + elif tokenH == 0x10: # int + return int.from_bytes(self._fp.read(1 << tokenL), 'big') + + elif token == 0x22: # real + return struct.unpack('>f', self._fp.read(4))[0] + + elif token == 0x23: # real + return struct.unpack('>d', self._fp.read(8))[0] + + elif token == 0x33: # date + f = struct.unpack('>d', self._fp.read(8))[0] + # timestamp 0 of binary plists corresponds to 1/1/2001 + # (year of Mac OS X 10.0), instead of 1/1/1970. + return datetime.datetime.utcfromtimestamp(f + (31 * 365 + 8) * 86400) + + elif tokenH == 0x40: # data + s = self._get_size(tokenL) + if self._use_builtin_types: + return self._fp.read(s) + else: + return Data(self._fp.read(s)) + + elif tokenH == 0x50: # ascii string + s = self._get_size(tokenL) + result = self._fp.read(s).decode('ascii') + return result + + elif tokenH == 0x60: # unicode string + s = self._get_size(tokenL) + return self._fp.read(s * 2).decode('utf-16be') + + # tokenH == 0x80 is documented as 'UID' and appears to be used for + # keyed-archiving, not in plists. + + elif tokenH == 0xA0: # array + s = self._get_size(tokenL) + obj_refs = self._read_refs(s) + return [self._read_object(self._object_offsets[x]) + for x in obj_refs] + + # tokenH == 0xB0 is documented as 'ordset', but is not actually + # implemented in the Apple reference code. + + # tokenH == 0xC0 is documented as 'set', but sets cannot be used in + # plists. + + elif tokenH == 0xD0: # dict + s = self._get_size(tokenL) + key_refs = self._read_refs(s) + obj_refs = self._read_refs(s) + result = self._dict_type() + for k, o in zip(key_refs, obj_refs): + result[self._read_object(self._object_offsets[k]) + ] = self._read_object(self._object_offsets[o]) + return result + + raise InvalidFileException() + +def _count_to_size(count): + if count < 1 << 8: + return 1 + + elif count < 1 << 16: + return 2 + + elif count << 1 << 32: + return 4 + + else: + return 8 + +class _BinaryPlistWriter (object): + def __init__(self, fp, sort_keys, skipkeys): + self._fp = fp + self._sort_keys = sort_keys + self._skipkeys = skipkeys + + def write(self, value): + + # Flattened object list: + self._objlist = [] + + # Mappings from object->objectid + # First dict has (type(object), object) as the key, + # second dict is used when object is not hashable and + # has id(object) as the key. + self._objtable = {} + self._objidtable = {} + + # Create list of all objects in the plist + self._flatten(value) + + # Size of object references in serialized containers + # depends on the number of objects in the plist. + num_objects = len(self._objlist) + self._object_offsets = [0]*num_objects + self._ref_size = _count_to_size(num_objects) + + self._ref_format = _BINARY_FORMAT[self._ref_size] + + # Write file header + self._fp.write(b'bplist00') + + # Write object list + for obj in self._objlist: + self._write_object(obj) + + # Write refnum->object offset table + top_object = self._getrefnum(value) + offset_table_offset = self._fp.tell() + offset_size = _count_to_size(offset_table_offset) + offset_format = '>' + _BINARY_FORMAT[offset_size] * num_objects + self._fp.write(struct.pack(offset_format, *self._object_offsets)) + + # Write trailer + sort_version = 0 + trailer = ( + sort_version, offset_size, self._ref_size, num_objects, + top_object, offset_table_offset + ) + self._fp.write(struct.pack('>5xBBBQQQ', *trailer)) + + def _flatten(self, value): + # First check if the object is in the object table, not used for + # containers to ensure that two subcontainers with the same contents + # will be serialized as distinct values. + if isinstance(value, ( + str, int, float, datetime.datetime, bytes, bytearray)): + if (type(value), value) in self._objtable: + return + + elif isinstance(value, Data): + if (type(value.data), value.data) in self._objtable: + return + + # Add to objectreference map + refnum = len(self._objlist) + self._objlist.append(value) + try: + if isinstance(value, Data): + self._objtable[(type(value.data), value.data)] = refnum + else: + self._objtable[(type(value), value)] = refnum + except TypeError: + self._objidtable[id(value)] = refnum + + # And finally recurse into containers + if isinstance(value, dict): + keys = [] + values = [] + items = value.items() + if self._sort_keys: + items = sorted(items) + + for k, v in items: + if not isinstance(k, str): + if self._skipkeys: + continue + raise TypeError("keys must be strings") + keys.append(k) + values.append(v) + + for o in itertools.chain(keys, values): + self._flatten(o) + + elif isinstance(value, (list, tuple)): + for o in value: + self._flatten(o) + + def _getrefnum(self, value): + try: + if isinstance(value, Data): + return self._objtable[(type(value.data), value.data)] + else: + return self._objtable[(type(value), value)] + except TypeError: + return self._objidtable[id(value)] + + def _write_size(self, token, size): + if size < 15: + self._fp.write(struct.pack('>B', token | size)) + + elif size < 1 << 8: + self._fp.write(struct.pack('>BBB', token | 0xF, 0x10, size)) + + elif size < 1 << 16: + self._fp.write(struct.pack('>BBH', token | 0xF, 0x11, size)) + + elif size < 1 << 32: + self._fp.write(struct.pack('>BBL', token | 0xF, 0x12, size)) + + else: + self._fp.write(struct.pack('>BBQ', token | 0xF, 0x13, size)) + + def _write_object(self, value): + ref = self._getrefnum(value) + self._object_offsets[ref] = self._fp.tell() + if value is None: + self._fp.write(b'\x00') + + elif value is False: + self._fp.write(b'\x08') + + elif value is True: + self._fp.write(b'\x09') + + elif isinstance(value, int): + if value < 1 << 8: + self._fp.write(struct.pack('>BB', 0x10, value)) + elif value < 1 << 16: + self._fp.write(struct.pack('>BH', 0x11, value)) + elif value < 1 << 32: + self._fp.write(struct.pack('>BL', 0x12, value)) + else: + self._fp.write(struct.pack('>BQ', 0x13, value)) + + elif isinstance(value, float): + self._fp.write(struct.pack('>Bd', 0x23, value)) + + elif isinstance(value, datetime.datetime): + f = (value - datetime.datetime(2001, 1, 1)).total_seconds() + self._fp.write(struct.pack('>Bd', 0x33, f)) + + elif isinstance(value, Data): + self._write_size(0x40, len(value.data)) + self._fp.write(value.data) + + elif isinstance(value, (bytes, bytearray)): + self._write_size(0x40, len(value)) + self._fp.write(value) + + elif isinstance(value, str): + try: + t = value.encode('ascii') + self._write_size(0x50, len(value)) + except UnicodeEncodeError: + t = value.encode('utf-16be') + self._write_size(0x60, len(value)) + + self._fp.write(t) + + elif isinstance(value, (list, tuple)): + refs = [self._getrefnum(o) for o in value] + s = len(refs) + self._write_size(0xA0, s) + self._fp.write(struct.pack('>' + self._ref_format * s, *refs)) + + elif isinstance(value, dict): + keyRefs, valRefs = [], [] + + if self._sort_keys: + rootItems = sorted(value.items()) + else: + rootItems = value.items() + + for k, v in rootItems: + if not isinstance(k, str): + if self._skipkeys: + continue + raise TypeError("keys must be strings") + keyRefs.append(self._getrefnum(k)) + valRefs.append(self._getrefnum(v)) + + s = len(keyRefs) + self._write_size(0xD0, s) + self._fp.write(struct.pack('>' + self._ref_format * s, *keyRefs)) + self._fp.write(struct.pack('>' + self._ref_format * s, *valRefs)) + + else: + raise InvalidFileException() + + +def _is_fmt_binary(header): + return header[:8] == b'bplist00' + + +# +# Generic bits +# + +_FORMATS={ + FMT_XML: dict( + detect=_is_fmt_xml, + parser=_PlistParser, + writer=_PlistWriter, + ), + FMT_BINARY: dict( + detect=_is_fmt_binary, + parser=_BinaryPlistParser, + writer=_BinaryPlistWriter, + ) +} + + +def load(fp, *, fmt=None, use_builtin_types=True, dict_type=dict): + """Read a .plist file. 'fp' should be (readable) file object. + Return the unpacked root object (which usually is a dictionary). + """ + if fmt is None: + header = fp.read(32) + fp.seek(0) + for info in _FORMATS.values(): + if info['detect'](header): + p = info['parser']( + use_builtin_types=use_builtin_types, + dict_type=dict_type, + ) + break + + else: + raise InvalidFileException() + + else: + p = _FORMATS[fmt]['parser'](use_builtin_types=use_builtin_types) + + return p.parse(fp) + + +def loads(value, *, fmt=None, use_builtin_types=True, dict_type=dict): + """Read a .plist file from a bytes object. + Return the unpacked root object (which usually is a dictionary). + """ + fp = BytesIO(value) + return load( + fp, fmt=fmt, use_builtin_types=use_builtin_types, dict_type=dict_type) + + +def dump(value, fp, *, fmt=FMT_XML, sort_keys=True, skipkeys=False): + """Write 'value' to a .plist file. 'fp' should be a (writable) + file object. + """ + if fmt not in _FORMATS: + raise ValueError("Unsupported format: %r"%(fmt,)) + + writer = _FORMATS[fmt]["writer"](fp, sort_keys=sort_keys, skipkeys=skipkeys) + writer.write(value) + + +def dumps(value, *, fmt=FMT_XML, skipkeys=False, sort_keys=True): + """Return a bytes object with the contents for a .plist file. + """ + fp = BytesIO() + dump(value, fp, fmt=fmt, skipkeys=skipkeys, sort_keys=sort_keys) + return fp.getvalue() -- cgit v1.2.1