X-Git-Url: http://www.git.cypherpunks.ru/?a=blobdiff_plain;f=pyderasn.py;h=3dd3ea9c151a3f1ca6f571a7765614e1c5bf5b91;hb=f928d1e65eeff56c2b5d511567e6b52417195f3b;hp=b9bc9dae515564a24dcc052005754d54993fca03;hpb=a34b3e75db1e13750560c89c2b80821f2e5d0d1f;p=pyderasn.git diff --git a/pyderasn.py b/pyderasn.py index b9bc9da..3dd3ea9 100755 --- a/pyderasn.py +++ b/pyderasn.py @@ -213,9 +213,11 @@ decoding process. Currently available context options: +* :ref:`allow_default_values ` +* :ref:`allow_expl_oob ` +* :ref:`allow_unordered_set ` * :ref:`bered ` * :ref:`defines_by_path ` -* :ref:`strict_default_existence ` .. _pprinting: @@ -274,7 +276,7 @@ You can specify multiple fields, that will be autodecoded -- that is why ``defines`` kwarg is a sequence. You can specify defined field relatively or absolutely to current decode path. For example ``defines`` for AlgorithmIdentifier of X.509's -``tbsCertificate.subjectPublicKeyInfo.algorithm.algorithm``:: +``tbsCertificate:subjectPublicKeyInfo:algorithm:algorithm``:: ( (("parameters",), { @@ -302,8 +304,7 @@ Following types can be automatically decoded (DEFINED BY): When any of those fields is automatically decoded, then ``.defined`` attribute contains ``(OID, value)`` tuple. ``OID`` tells by which OID it was defined, ``value`` contains corresponding decoded value. For example -above, ``content_info["content"].defined == (id_signedData, -signed_data)``. +above, ``content_info["content"].defined == (id_signedData, signed_data)``. .. _defines_by_path_ctx: @@ -383,7 +384,7 @@ constructed primitive types should be parsed successfully. * If object is encoded in BER form (not the DER one), then ``bered`` attribute is set to True. Only ``BOOLEAN``, ``BIT STRING``, ``OCTET - STRING`` can contain it. + STRING``, ``SEQUENCE``, ``SET`` can contain it. * If object has an indefinite length encoding, then its ``lenindef`` attribute is set to True. Only ``BIT STRING``, ``OCTET STRING``, ``SEQUENCE``, ``SET``, ``SEQUENCE OF``, ``SET OF``, ``ANY`` can @@ -394,6 +395,22 @@ constructed primitive types should be parsed successfully. EOC (end-of-contents) token's length is taken in advance in object's value length. +.. _allow_expl_oob_ctx: + +Allow explicit tag out-of-bound +------------------------------- + +Invalid BER encoding could contain ``EXPLICIT`` tag containing more than +one value, more than one object. If you set ``allow_expl_oob`` context +option to True, then no error will be raised and that invalid encoding +will be silently further processed. But pay attention that offsets and +lengths will be invalid in that case. + +.. warning:: + + This option should be used only for skipping some decode errors, just + to see the decoded structure somehow. + Primitive types --------------- @@ -641,7 +658,7 @@ class DecodeError(Exception): c for c in ( "" if self.klass is None else self.klass.__name__, ( - ("(%s)" % ".".join(str(dp) for dp in self.decode_path)) + ("(%s)" % ":".join(str(dp) for dp in self.decode_path)) if len(self.decode_path) > 0 else "" ), ("(at %d)" % self.offset) if self.offset > 0 else "", @@ -1093,6 +1110,13 @@ class Obj(object): if tag_only: return obj, tail = result + if obj.tlvlen < l and not ctx.get("allow_expl_oob", False): + raise DecodeError( + "explicit tag out-of-bound, longer than data", + klass=self.__class__, + decode_path=decode_path, + offset=offset, + ) return obj, (tail if leavemm else tail.tobytes()) @property @@ -1258,7 +1282,7 @@ def _pp( ) -def _colorize(what, colour, with_colours, attrs=("bold",)): +def _colourize(what, colour, with_colours, attrs=("bold",)): return colored(what, colour, attrs=attrs) if with_colours else what @@ -1268,6 +1292,8 @@ def pp_console_row( with_offsets=False, with_blob=True, with_colours=False, + with_decode_path=False, + decode_path_len_decrease=0, ): cols = [] if with_offsets: @@ -1279,20 +1305,21 @@ def pp_console_row( ), LENINDEF_PP_CHAR if pp.expl_lenindef else " ", ) - cols.append(_colorize(col, "red", with_colours, ())) + cols.append(_colourize(col, "red", with_colours, ())) col = "[%d,%d,%4d]%s" % ( pp.tlen, pp.llen, pp.vlen, LENINDEF_PP_CHAR if pp.lenindef else " " ) - col = _colorize(col, "green", with_colours, ()) + col = _colourize(col, "green", with_colours, ()) cols.append(col) - if len(pp.decode_path) > 0: - cols.append(" ." * (len(pp.decode_path))) + decode_path_len = len(pp.decode_path) - decode_path_len_decrease + if decode_path_len > 0: + cols.append(" ." * decode_path_len) ent = pp.decode_path[-1] if isinstance(ent, DecodePathDefBy): - cols.append(_colorize("DEFINED BY", "red", with_colours, ("reverse",))) + cols.append(_colourize("DEFINED BY", "red", with_colours, ("reverse",))) value = str(ent.defined_by) if ( oids is not None and @@ -1300,49 +1327,56 @@ def pp_console_row( ObjectIdentifier.asn1_type_name and value in oids ): - cols.append(_colorize("%s:" % oids[value], "green", with_colours)) + cols.append(_colourize("%s:" % oids[value], "green", with_colours)) else: - cols.append(_colorize("%s:" % value, "white", with_colours, ("reverse",))) + cols.append(_colourize("%s:" % value, "white", with_colours, ("reverse",))) else: - cols.append(_colorize("%s:" % ent, "yellow", with_colours, ("reverse",))) + cols.append(_colourize("%s:" % ent, "yellow", with_colours, ("reverse",))) if pp.expl is not None: klass, _, num = pp.expl col = "[%s%d] EXPLICIT" % (TagClassReprs[klass], num) - cols.append(_colorize(col, "blue", with_colours)) + cols.append(_colourize(col, "blue", with_colours)) if pp.impl is not None: klass, _, num = pp.impl col = "[%s%d]" % (TagClassReprs[klass], num) - cols.append(_colorize(col, "blue", with_colours)) + cols.append(_colourize(col, "blue", with_colours)) if pp.asn1_type_name.replace(" ", "") != pp.obj_name.upper(): - cols.append(_colorize(pp.obj_name, "magenta", with_colours)) + cols.append(_colourize(pp.obj_name, "magenta", with_colours)) if pp.bered: - cols.append(_colorize("BER", "red", with_colours)) - cols.append(_colorize(pp.asn1_type_name, "cyan", with_colours)) + cols.append(_colourize("BER", "red", with_colours)) + cols.append(_colourize(pp.asn1_type_name, "cyan", with_colours)) if pp.value is not None: value = pp.value - cols.append(_colorize(value, "white", with_colours, ("reverse",))) + cols.append(_colourize(value, "white", with_colours, ("reverse",))) if ( oids is not None and pp.asn1_type_name == ObjectIdentifier.asn1_type_name and value in oids ): - cols.append(_colorize("(%s)" % oids[value], "green", with_colours)) + cols.append(_colourize("(%s)" % oids[value], "green", with_colours)) if with_blob: if isinstance(pp.blob, binary_type): cols.append(hexenc(pp.blob)) elif isinstance(pp.blob, tuple): cols.append(", ".join(pp.blob)) if pp.optional: - cols.append(_colorize("OPTIONAL", "red", with_colours)) + cols.append(_colourize("OPTIONAL", "red", with_colours)) if pp.default: - cols.append(_colorize("DEFAULT", "red", with_colours)) + cols.append(_colourize("DEFAULT", "red", with_colours)) + if with_decode_path: + cols.append(_colourize( + "[%s]" % ":".join(str(p) for p in pp.decode_path), + "grey", + with_colours, + )) return " ".join(cols) -def pp_console_blob(pp): +def pp_console_blob(pp, decode_path_len_decrease=0): cols = [" " * len("XXXXXYYZ [X,X,XXXX]Z")] - if len(pp.decode_path) > 0: - cols.append(" ." * (len(pp.decode_path) + 1)) + decode_path_len = len(pp.decode_path) - decode_path_len_decrease + if decode_path_len > 0: + cols.append(" ." * (decode_path_len + 1)) if isinstance(pp.blob, binary_type): blob = hexenc(pp.blob).upper() for i in range(0, len(blob), 32): @@ -1354,7 +1388,14 @@ def pp_console_blob(pp): yield " ".join(cols + [", ".join(pp.blob)]) -def pprint(obj, oids=None, big_blobs=False, with_colours=False): +def pprint( + obj, + oids=None, + big_blobs=False, + with_colours=False, + with_decode_path=False, + decode_path_only=(), +): """Pretty print object :param Obj obj: object you want to pretty print @@ -1365,10 +1406,19 @@ def pprint(obj, oids=None, big_blobs=False, with_colours=False): lines :param with_colours: colourize output, if ``termcolor`` library is available + :param with_decode_path: print decode path + :param decode_path_only: print only that specified decode path """ def _pprint_pps(pps): for pp in pps: if hasattr(pp, "_fields"): + if ( + decode_path_only != () and + tuple( + str(p) for p in pp.decode_path[:len(decode_path_only)] + ) != decode_path_only + ): + continue if big_blobs: yield pp_console_row( pp, @@ -1376,8 +1426,13 @@ def pprint(obj, oids=None, big_blobs=False, with_colours=False): with_offsets=True, with_blob=False, with_colours=with_colours, + with_decode_path=with_decode_path, + decode_path_len_decrease=len(decode_path_only), ) - for row in pp_console_blob(pp): + for row in pp_console_blob( + pp, + decode_path_len_decrease=len(decode_path_only), + ): yield row else: yield pp_console_row( @@ -1386,6 +1441,8 @@ def pprint(obj, oids=None, big_blobs=False, with_colours=False): with_offsets=True, with_blob=True, with_colours=with_colours, + with_decode_path=with_decode_path, + decode_path_len_decrease=len(decode_path_only), ) else: for row in _pprint_pps(pp): @@ -4282,18 +4339,14 @@ class Sequence(Obj): All defaulted values are always optional. - .. _strict_default_existence_ctx: + .. _allow_default_values_ctx: - .. warning:: - - When decoded DER contains defaulted value inside, then - technically this is not valid DER encoding. But we allow and pass - it **by default**. Of course reencoding of that kind of DER will - result in different binary representation (validly without - defaulted value inside). You can enable strict defaulted values - existence validation by setting ``"strict_default_existence": - True`` :ref:`context ` option -- decoding process will raise - an exception if defaulted value is met. + DER prohibits default value encoding and will raise an error if + default value is unexpectedly met during decode. + If :ref:`bered ` context option is set, then no error + will be raised, but ``bered`` attribute set. You can disable strict + defaulted values existence validation by setting + ``"allow_default_values": True`` :ref:`context ` option. Two sequences are equal if they have equal specification (schema), implicit/explicit tagging and the same values. @@ -4453,10 +4506,11 @@ class Sequence(Obj): if tag_only: return lenindef = False + ctx_bered = ctx.get("bered", False) try: l, llen, v = len_decode(lv) except LenIndefForm as err: - if not ctx.get("bered", False): + if not ctx_bered: raise err.__class__( msg=err.msg, klass=self.__class__, @@ -4484,6 +4538,8 @@ class Sequence(Obj): vlen = 0 sub_offset = offset + tlen + llen values = {} + bered = False + ctx_allow_default_values = ctx.get("allow_default_values", False) for name, spec in self.specs.items(): if spec.optional and ( (lenindef and v[:EOC_LEN].tobytes() == EOC) or @@ -4556,15 +4612,15 @@ class Sequence(Obj): sub_offset += value_len v = v_tail if spec.default is not None and value == spec.default: - if ctx.get("strict_default_existence", False): + if ctx_bered or ctx_allow_default_values: + bered = True + else: raise DecodeError( "DEFAULT value met", klass=self.__class__, decode_path=sub_decode_path, offset=sub_offset, ) - else: - continue values[name] = value spec_defines = getattr(spec, "defines", ()) @@ -4607,6 +4663,7 @@ class Sequence(Obj): ) obj._value = values obj.lenindef = lenindef + obj.bered = bered return obj, tail def __repr__(self): @@ -4652,6 +4709,14 @@ class Set(Sequence): """``SET`` structure type Its usage is identical to :py:class:`pyderasn.Sequence`. + + .. _allow_unordered_set_ctx: + + DER prohibits unordered values encoding and will raise an error + during decode. If If :ref:`bered ` context option is set, + then no error will occure. Also you can disable strict values + ordering check by setting ``"allow_unordered_set": True`` + :ref:`context ` option. """ __slots__ = () tag_default = tag_encode(form=TagFormConstructed, num=17) @@ -4682,10 +4747,11 @@ class Set(Sequence): if tag_only: return lenindef = False + ctx_bered = ctx.get("bered", False) try: l, llen, v = len_decode(lv) except LenIndefForm as err: - if not ctx.get("bered", False): + if not ctx_bered: raise err.__class__( msg=err.msg, klass=self.__class__, @@ -4712,6 +4778,10 @@ class Set(Sequence): vlen = 0 sub_offset = offset + tlen + llen values = {} + bered = False + ctx_allow_default_values = ctx.get("allow_default_values", False) + ctx_allow_unordered_set = ctx.get("allow_unordered_set", False) + value_prev = memoryview(v[:0]) specs_items = self.specs.items while len(v) > 0: if lenindef and v[:EOC_LEN].tobytes() == EOC: @@ -4744,12 +4814,32 @@ class Set(Sequence): ctx=ctx, ) value_len = value.fulllen + if value_prev.tobytes() > v[:value_len].tobytes(): + if ctx_bered or ctx_allow_unordered_set: + bered = True + else: + raise DecodeError( + "unordered " + self.asn1_type_name, + klass=self.__class__, + decode_path=sub_decode_path, + offset=sub_offset, + ) + if spec.default is None or value != spec.default: + pass + elif ctx_bered or ctx_allow_default_values: + bered = True + else: + raise DecodeError( + "DEFAULT value met", + klass=self.__class__, + decode_path=sub_decode_path, + offset=sub_offset, + ) + values[name] = value + value_prev = v[:value_len] sub_offset += value_len vlen += value_len v = v_tail - if spec.default is None or value != spec.default: # pragma: no cover - # SeqMixing.test_encoded_default_accepted covers that place - values[name] = value obj = self.__class__( schema=self.specs, impl=self.tag, @@ -4758,7 +4848,6 @@ class Set(Sequence): optional=self.optional, _decoded=(offset, llen, vlen + (EOC_LEN if lenindef else 0)), ) - obj._value = values if lenindef: if v[:EOC_LEN].tobytes() != EOC: raise DecodeError( @@ -4769,6 +4858,7 @@ class Set(Sequence): ) tail = v[EOC_LEN:] obj.lenindef = True + obj._value = values if not obj.ready: raise DecodeError( "not all values are ready", @@ -4776,6 +4866,7 @@ class Set(Sequence): decode_path=decode_path, offset=offset, ) + obj.bered = bered return obj, tail @@ -5025,16 +5116,24 @@ class SequenceOf(Obj): vlen += value_len v = v_tail _value.append(value) - obj = self.__class__( - value=_value, - schema=spec, - bounds=(self._bound_min, self._bound_max), - impl=self.tag, - expl=self._expl, - default=self.default, - optional=self.optional, - _decoded=(offset, llen, vlen + (EOC_LEN if lenindef else 0)), - ) + try: + obj = self.__class__( + value=_value, + schema=spec, + bounds=(self._bound_min, self._bound_max), + impl=self.tag, + expl=self._expl, + default=self.default, + optional=self.optional, + _decoded=(offset, llen, vlen + (EOC_LEN if lenindef else 0)), + ) + except BoundsError as err: + raise DecodeError( + msg=str(err), + klass=self.__class__, + decode_path=decode_path, + offset=offset, + ) if lenindef: if v[:EOC_LEN].tobytes() != EOC: raise DecodeError( @@ -5130,10 +5229,21 @@ def generic_decoder(): # pragma: no cover __slots__ = () schema = choice - def pprint_any(obj, oids=None, with_colours=False): + def pprint_any( + obj, + oids=None, + with_colours=False, + with_decode_path=False, + decode_path_only=(), + ): def _pprint_pps(pps): for pp in pps: if hasattr(pp, "_fields"): + if ( + decode_path_only != () and + pp.decode_path[:len(decode_path_only)] != decode_path_only + ): + continue if pp.asn1_type_name == Choice.asn1_type_name: continue pp_kwargs = pp._asdict() @@ -5145,8 +5255,13 @@ def generic_decoder(): # pragma: no cover with_offsets=True, with_blob=False, with_colours=with_colours, + with_decode_path=with_decode_path, + decode_path_len_decrease=len(decode_path_only), ) - for row in pp_console_blob(pp): + for row in pp_console_blob( + pp, + decode_path_len_decrease=len(decode_path_only), + ): yield row else: for row in _pprint_pps(pp): @@ -5178,9 +5293,23 @@ def main(): # pragma: no cover ) parser.add_argument( "--nobered", - action='store_true', + action="store_true", help="Disallow BER encoding", ) + parser.add_argument( + "--print-decode-path", + action="store_true", + help="Print decode paths", + ) + parser.add_argument( + "--decode-path-only", + help="Print only specified decode path", + ) + parser.add_argument( + "--allow-expl-oob", + action="store_true", + help="Allow explicit tag out-of-bound", + ) parser.add_argument( "DERFile", type=argparse.FileType("rb"), @@ -5197,7 +5326,10 @@ def main(): # pragma: no cover pprinter = partial(pprint, big_blobs=True) else: schema, pprinter = generic_decoder() - ctx = {"bered": not args.nobered} + ctx = { + "bered": not args.nobered, + "allow_expl_oob": args.allow_expl_oob, + } if args.defines_by_path is not None: ctx["defines_by_path"] = obj_by_path(args.defines_by_path) obj, tail = schema().decode(der, ctx=ctx) @@ -5205,6 +5337,11 @@ def main(): # pragma: no cover obj, oids=oids, with_colours=True if environ.get("NO_COLOR") is None else False, + with_decode_path=args.print_decode_path, + decode_path_only=( + () if args.decode_path_only is None else + tuple(args.decode_path_only.split(":")) + ), )) if tail != b"": print("\nTrailing data: %s" % hexenc(tail))