You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
517 lines
16 KiB
517 lines
16 KiB
# Copyright (C) Dnspython Contributors, see LICENSE for text of ISC license
|
|
|
|
# Copyright (C) 2001-2017 Nominum, Inc.
|
|
#
|
|
# Permission to use, copy, modify, and distribute this software and its
|
|
# documentation for any purpose with or without fee is hereby granted,
|
|
# provided that the above copyright notice and this permission notice
|
|
# appear in all copies.
|
|
#
|
|
# THE SOFTWARE IS PROVIDED "AS IS" AND NOMINUM DISCLAIMS ALL WARRANTIES
|
|
# WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
|
|
# MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL NOMINUM BE LIABLE FOR
|
|
# ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
|
|
# WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
|
|
# ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT
|
|
# OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
|
|
|
|
"""DNS rdatasets (an rdataset is a set of rdatas of a given type and class)"""
|
|
|
|
import io
|
|
import random
|
|
import struct
|
|
from typing import Any, Collection, Dict, List, Optional, Union, cast
|
|
|
|
import dns.exception
|
|
import dns.immutable
|
|
import dns.name
|
|
import dns.rdata
|
|
import dns.rdataclass
|
|
import dns.rdatatype
|
|
import dns.renderer
|
|
import dns.set
|
|
import dns.ttl
|
|
|
|
# define SimpleSet here for backwards compatibility
|
|
SimpleSet = dns.set.Set
|
|
|
|
|
|
class DifferingCovers(dns.exception.DNSException):
|
|
"""An attempt was made to add a DNS SIG/RRSIG whose covered type
|
|
is not the same as that of the other rdatas in the rdataset."""
|
|
|
|
|
|
class IncompatibleTypes(dns.exception.DNSException):
|
|
"""An attempt was made to add DNS RR data of an incompatible type."""
|
|
|
|
|
|
class Rdataset(dns.set.Set):
|
|
"""A DNS rdataset."""
|
|
|
|
__slots__ = ["rdclass", "rdtype", "covers", "ttl"]
|
|
|
|
def __init__(
|
|
self,
|
|
rdclass: dns.rdataclass.RdataClass,
|
|
rdtype: dns.rdatatype.RdataType,
|
|
covers: dns.rdatatype.RdataType = dns.rdatatype.NONE,
|
|
ttl: int = 0,
|
|
):
|
|
"""Create a new rdataset of the specified class and type.
|
|
|
|
*rdclass*, a ``dns.rdataclass.RdataClass``, the rdataclass.
|
|
|
|
*rdtype*, an ``dns.rdatatype.RdataType``, the rdatatype.
|
|
|
|
*covers*, an ``dns.rdatatype.RdataType``, the covered rdatatype.
|
|
|
|
*ttl*, an ``int``, the TTL.
|
|
"""
|
|
|
|
super().__init__()
|
|
self.rdclass = rdclass
|
|
self.rdtype: dns.rdatatype.RdataType = rdtype
|
|
self.covers: dns.rdatatype.RdataType = covers
|
|
self.ttl = ttl
|
|
|
|
def _clone(self):
|
|
obj = super()._clone()
|
|
obj.rdclass = self.rdclass
|
|
obj.rdtype = self.rdtype
|
|
obj.covers = self.covers
|
|
obj.ttl = self.ttl
|
|
return obj
|
|
|
|
def update_ttl(self, ttl: int) -> None:
|
|
"""Perform TTL minimization.
|
|
|
|
Set the TTL of the rdataset to be the lesser of the set's current
|
|
TTL or the specified TTL. If the set contains no rdatas, set the TTL
|
|
to the specified TTL.
|
|
|
|
*ttl*, an ``int`` or ``str``.
|
|
"""
|
|
ttl = dns.ttl.make(ttl)
|
|
if len(self) == 0:
|
|
self.ttl = ttl
|
|
elif ttl < self.ttl:
|
|
self.ttl = ttl
|
|
|
|
def add( # pylint: disable=arguments-differ,arguments-renamed
|
|
self, rd: dns.rdata.Rdata, ttl: Optional[int] = None
|
|
) -> None:
|
|
"""Add the specified rdata to the rdataset.
|
|
|
|
If the optional *ttl* parameter is supplied, then
|
|
``self.update_ttl(ttl)`` will be called prior to adding the rdata.
|
|
|
|
*rd*, a ``dns.rdata.Rdata``, the rdata
|
|
|
|
*ttl*, an ``int``, the TTL.
|
|
|
|
Raises ``dns.rdataset.IncompatibleTypes`` if the type and class
|
|
do not match the type and class of the rdataset.
|
|
|
|
Raises ``dns.rdataset.DifferingCovers`` if the type is a signature
|
|
type and the covered type does not match that of the rdataset.
|
|
"""
|
|
|
|
#
|
|
# If we're adding a signature, do some special handling to
|
|
# check that the signature covers the same type as the
|
|
# other rdatas in this rdataset. If this is the first rdata
|
|
# in the set, initialize the covers field.
|
|
#
|
|
if self.rdclass != rd.rdclass or self.rdtype != rd.rdtype:
|
|
raise IncompatibleTypes
|
|
if ttl is not None:
|
|
self.update_ttl(ttl)
|
|
if self.rdtype == dns.rdatatype.RRSIG or self.rdtype == dns.rdatatype.SIG:
|
|
covers = rd.covers()
|
|
if len(self) == 0 and self.covers == dns.rdatatype.NONE:
|
|
self.covers = covers
|
|
elif self.covers != covers:
|
|
raise DifferingCovers
|
|
if dns.rdatatype.is_singleton(rd.rdtype) and len(self) > 0:
|
|
self.clear()
|
|
super().add(rd)
|
|
|
|
def union_update(self, other):
|
|
self.update_ttl(other.ttl)
|
|
super().union_update(other)
|
|
|
|
def intersection_update(self, other):
|
|
self.update_ttl(other.ttl)
|
|
super().intersection_update(other)
|
|
|
|
def update(self, other):
|
|
"""Add all rdatas in other to self.
|
|
|
|
*other*, a ``dns.rdataset.Rdataset``, the rdataset from which
|
|
to update.
|
|
"""
|
|
|
|
self.update_ttl(other.ttl)
|
|
super().update(other)
|
|
|
|
def _rdata_repr(self):
|
|
def maybe_truncate(s):
|
|
if len(s) > 100:
|
|
return s[:100] + "..."
|
|
return s
|
|
|
|
return "[%s]" % ", ".join("<%s>" % maybe_truncate(str(rr)) for rr in self)
|
|
|
|
def __repr__(self):
|
|
if self.covers == 0:
|
|
ctext = ""
|
|
else:
|
|
ctext = "(" + dns.rdatatype.to_text(self.covers) + ")"
|
|
return (
|
|
"<DNS "
|
|
+ dns.rdataclass.to_text(self.rdclass)
|
|
+ " "
|
|
+ dns.rdatatype.to_text(self.rdtype)
|
|
+ ctext
|
|
+ " rdataset: "
|
|
+ self._rdata_repr()
|
|
+ ">"
|
|
)
|
|
|
|
def __str__(self):
|
|
return self.to_text()
|
|
|
|
def __eq__(self, other):
|
|
if not isinstance(other, Rdataset):
|
|
return False
|
|
if (
|
|
self.rdclass != other.rdclass
|
|
or self.rdtype != other.rdtype
|
|
or self.covers != other.covers
|
|
):
|
|
return False
|
|
return super().__eq__(other)
|
|
|
|
def __ne__(self, other):
|
|
return not self.__eq__(other)
|
|
|
|
def to_text(
|
|
self,
|
|
name: Optional[dns.name.Name] = None,
|
|
origin: Optional[dns.name.Name] = None,
|
|
relativize: bool = True,
|
|
override_rdclass: Optional[dns.rdataclass.RdataClass] = None,
|
|
want_comments: bool = False,
|
|
**kw: Dict[str, Any],
|
|
) -> str:
|
|
"""Convert the rdataset into DNS zone file format.
|
|
|
|
See ``dns.name.Name.choose_relativity`` for more information
|
|
on how *origin* and *relativize* determine the way names
|
|
are emitted.
|
|
|
|
Any additional keyword arguments are passed on to the rdata
|
|
``to_text()`` method.
|
|
|
|
*name*, a ``dns.name.Name``. If name is not ``None``, emit RRs with
|
|
*name* as the owner name.
|
|
|
|
*origin*, a ``dns.name.Name`` or ``None``, the origin for relative
|
|
names.
|
|
|
|
*relativize*, a ``bool``. If ``True``, names will be relativized
|
|
to *origin*.
|
|
|
|
*override_rdclass*, a ``dns.rdataclass.RdataClass`` or ``None``.
|
|
If not ``None``, use this class instead of the Rdataset's class.
|
|
|
|
*want_comments*, a ``bool``. If ``True``, emit comments for rdata
|
|
which have them. The default is ``False``.
|
|
"""
|
|
|
|
if name is not None:
|
|
name = name.choose_relativity(origin, relativize)
|
|
ntext = str(name)
|
|
pad = " "
|
|
else:
|
|
ntext = ""
|
|
pad = ""
|
|
s = io.StringIO()
|
|
if override_rdclass is not None:
|
|
rdclass = override_rdclass
|
|
else:
|
|
rdclass = self.rdclass
|
|
if len(self) == 0:
|
|
#
|
|
# Empty rdatasets are used for the question section, and in
|
|
# some dynamic updates, so we don't need to print out the TTL
|
|
# (which is meaningless anyway).
|
|
#
|
|
s.write(
|
|
"{}{}{} {}\n".format(
|
|
ntext,
|
|
pad,
|
|
dns.rdataclass.to_text(rdclass),
|
|
dns.rdatatype.to_text(self.rdtype),
|
|
)
|
|
)
|
|
else:
|
|
for rd in self:
|
|
extra = ""
|
|
if want_comments:
|
|
if rd.rdcomment:
|
|
extra = f" ;{rd.rdcomment}"
|
|
s.write(
|
|
"%s%s%d %s %s %s%s\n"
|
|
% (
|
|
ntext,
|
|
pad,
|
|
self.ttl,
|
|
dns.rdataclass.to_text(rdclass),
|
|
dns.rdatatype.to_text(self.rdtype),
|
|
rd.to_text(origin=origin, relativize=relativize, **kw),
|
|
extra,
|
|
)
|
|
)
|
|
#
|
|
# We strip off the final \n for the caller's convenience in printing
|
|
#
|
|
return s.getvalue()[:-1]
|
|
|
|
def to_wire(
|
|
self,
|
|
name: dns.name.Name,
|
|
file: Any,
|
|
compress: Optional[dns.name.CompressType] = None,
|
|
origin: Optional[dns.name.Name] = None,
|
|
override_rdclass: Optional[dns.rdataclass.RdataClass] = None,
|
|
want_shuffle: bool = True,
|
|
) -> int:
|
|
"""Convert the rdataset to wire format.
|
|
|
|
*name*, a ``dns.name.Name`` is the owner name to use.
|
|
|
|
*file* is the file where the name is emitted (typically a
|
|
BytesIO file).
|
|
|
|
*compress*, a ``dict``, is the compression table to use. If
|
|
``None`` (the default), names will not be compressed.
|
|
|
|
*origin* is a ``dns.name.Name`` or ``None``. If the name is
|
|
relative and origin is not ``None``, then *origin* will be appended
|
|
to it.
|
|
|
|
*override_rdclass*, an ``int``, is used as the class instead of the
|
|
class of the rdataset. This is useful when rendering rdatasets
|
|
associated with dynamic updates.
|
|
|
|
*want_shuffle*, a ``bool``. If ``True``, then the order of the
|
|
Rdatas within the Rdataset will be shuffled before rendering.
|
|
|
|
Returns an ``int``, the number of records emitted.
|
|
"""
|
|
|
|
if override_rdclass is not None:
|
|
rdclass = override_rdclass
|
|
want_shuffle = False
|
|
else:
|
|
rdclass = self.rdclass
|
|
if len(self) == 0:
|
|
name.to_wire(file, compress, origin)
|
|
file.write(struct.pack("!HHIH", self.rdtype, rdclass, 0, 0))
|
|
return 1
|
|
else:
|
|
l: Union[Rdataset, List[dns.rdata.Rdata]]
|
|
if want_shuffle:
|
|
l = list(self)
|
|
random.shuffle(l)
|
|
else:
|
|
l = self
|
|
for rd in l:
|
|
name.to_wire(file, compress, origin)
|
|
file.write(struct.pack("!HHI", self.rdtype, rdclass, self.ttl))
|
|
with dns.renderer.prefixed_length(file, 2):
|
|
rd.to_wire(file, compress, origin)
|
|
return len(self)
|
|
|
|
def match(
|
|
self,
|
|
rdclass: dns.rdataclass.RdataClass,
|
|
rdtype: dns.rdatatype.RdataType,
|
|
covers: dns.rdatatype.RdataType,
|
|
) -> bool:
|
|
"""Returns ``True`` if this rdataset matches the specified class,
|
|
type, and covers.
|
|
"""
|
|
if self.rdclass == rdclass and self.rdtype == rdtype and self.covers == covers:
|
|
return True
|
|
return False
|
|
|
|
def processing_order(self) -> List[dns.rdata.Rdata]:
|
|
"""Return rdatas in a valid processing order according to the type's
|
|
specification. For example, MX records are in preference order from
|
|
lowest to highest preferences, with items of the same preference
|
|
shuffled.
|
|
|
|
For types that do not define a processing order, the rdatas are
|
|
simply shuffled.
|
|
"""
|
|
if len(self) == 0:
|
|
return []
|
|
else:
|
|
return self[0]._processing_order(iter(self))
|
|
|
|
|
|
@dns.immutable.immutable
|
|
class ImmutableRdataset(Rdataset): # lgtm[py/missing-equals]
|
|
"""An immutable DNS rdataset."""
|
|
|
|
_clone_class = Rdataset
|
|
|
|
def __init__(self, rdataset: Rdataset):
|
|
"""Create an immutable rdataset from the specified rdataset."""
|
|
|
|
super().__init__(
|
|
rdataset.rdclass, rdataset.rdtype, rdataset.covers, rdataset.ttl
|
|
)
|
|
self.items = dns.immutable.Dict(rdataset.items)
|
|
|
|
def update_ttl(self, ttl):
|
|
raise TypeError("immutable")
|
|
|
|
def add(self, rd, ttl=None):
|
|
raise TypeError("immutable")
|
|
|
|
def union_update(self, other):
|
|
raise TypeError("immutable")
|
|
|
|
def intersection_update(self, other):
|
|
raise TypeError("immutable")
|
|
|
|
def update(self, other):
|
|
raise TypeError("immutable")
|
|
|
|
def __delitem__(self, i):
|
|
raise TypeError("immutable")
|
|
|
|
# lgtm complains about these not raising ArithmeticError, but there is
|
|
# precedent for overrides of these methods in other classes to raise
|
|
# TypeError, and it seems like the better exception.
|
|
|
|
def __ior__(self, other): # lgtm[py/unexpected-raise-in-special-method]
|
|
raise TypeError("immutable")
|
|
|
|
def __iand__(self, other): # lgtm[py/unexpected-raise-in-special-method]
|
|
raise TypeError("immutable")
|
|
|
|
def __iadd__(self, other): # lgtm[py/unexpected-raise-in-special-method]
|
|
raise TypeError("immutable")
|
|
|
|
def __isub__(self, other): # lgtm[py/unexpected-raise-in-special-method]
|
|
raise TypeError("immutable")
|
|
|
|
def clear(self):
|
|
raise TypeError("immutable")
|
|
|
|
def __copy__(self):
|
|
return ImmutableRdataset(super().copy())
|
|
|
|
def copy(self):
|
|
return ImmutableRdataset(super().copy())
|
|
|
|
def union(self, other):
|
|
return ImmutableRdataset(super().union(other))
|
|
|
|
def intersection(self, other):
|
|
return ImmutableRdataset(super().intersection(other))
|
|
|
|
def difference(self, other):
|
|
return ImmutableRdataset(super().difference(other))
|
|
|
|
def symmetric_difference(self, other):
|
|
return ImmutableRdataset(super().symmetric_difference(other))
|
|
|
|
|
|
def from_text_list(
|
|
rdclass: Union[dns.rdataclass.RdataClass, str],
|
|
rdtype: Union[dns.rdatatype.RdataType, str],
|
|
ttl: int,
|
|
text_rdatas: Collection[str],
|
|
idna_codec: Optional[dns.name.IDNACodec] = None,
|
|
origin: Optional[dns.name.Name] = None,
|
|
relativize: bool = True,
|
|
relativize_to: Optional[dns.name.Name] = None,
|
|
) -> Rdataset:
|
|
"""Create an rdataset with the specified class, type, and TTL, and with
|
|
the specified list of rdatas in text format.
|
|
|
|
*idna_codec*, a ``dns.name.IDNACodec``, specifies the IDNA
|
|
encoder/decoder to use; if ``None``, the default IDNA 2003
|
|
encoder/decoder is used.
|
|
|
|
*origin*, a ``dns.name.Name`` (or ``None``), the
|
|
origin to use for relative names.
|
|
|
|
*relativize*, a ``bool``. If true, name will be relativized.
|
|
|
|
*relativize_to*, a ``dns.name.Name`` (or ``None``), the origin to use
|
|
when relativizing names. If not set, the *origin* value will be used.
|
|
|
|
Returns a ``dns.rdataset.Rdataset`` object.
|
|
"""
|
|
|
|
rdclass = dns.rdataclass.RdataClass.make(rdclass)
|
|
rdtype = dns.rdatatype.RdataType.make(rdtype)
|
|
r = Rdataset(rdclass, rdtype)
|
|
r.update_ttl(ttl)
|
|
for t in text_rdatas:
|
|
rd = dns.rdata.from_text(
|
|
r.rdclass, r.rdtype, t, origin, relativize, relativize_to, idna_codec
|
|
)
|
|
r.add(rd)
|
|
return r
|
|
|
|
|
|
def from_text(
|
|
rdclass: Union[dns.rdataclass.RdataClass, str],
|
|
rdtype: Union[dns.rdatatype.RdataType, str],
|
|
ttl: int,
|
|
*text_rdatas: Any,
|
|
) -> Rdataset:
|
|
"""Create an rdataset with the specified class, type, and TTL, and with
|
|
the specified rdatas in text format.
|
|
|
|
Returns a ``dns.rdataset.Rdataset`` object.
|
|
"""
|
|
|
|
return from_text_list(rdclass, rdtype, ttl, cast(Collection[str], text_rdatas))
|
|
|
|
|
|
def from_rdata_list(ttl: int, rdatas: Collection[dns.rdata.Rdata]) -> Rdataset:
|
|
"""Create an rdataset with the specified TTL, and with
|
|
the specified list of rdata objects.
|
|
|
|
Returns a ``dns.rdataset.Rdataset`` object.
|
|
"""
|
|
|
|
if len(rdatas) == 0:
|
|
raise ValueError("rdata list must not be empty")
|
|
r = None
|
|
for rd in rdatas:
|
|
if r is None:
|
|
r = Rdataset(rd.rdclass, rd.rdtype)
|
|
r.update_ttl(ttl)
|
|
r.add(rd)
|
|
assert r is not None
|
|
return r
|
|
|
|
|
|
def from_rdata(ttl: int, *rdatas: Any) -> Rdataset:
|
|
"""Create an rdataset with the specified TTL, and with
|
|
the specified rdata objects.
|
|
|
|
Returns a ``dns.rdataset.Rdataset`` object.
|
|
"""
|
|
|
|
return from_rdata_list(ttl, cast(Collection[dns.rdata.Rdata], rdatas))
|