Support multiple 'id' lines in microdescriptors

With the ed25519 additions the 'id' lines of microdescriptors can now appear multiple times... https://gitweb.torproject.org/torspec.git/commit/?id=09ff9e2

Support multiple 'id' lines in microdescriptors
373d56f8 · Damian Johnson · 66e39330 · 373d56f8 · 373d56f8 · 373d56f8
Commit 373d56f8 authored Aug 26, 2015 by Damian Johnson
--- a/docs/change_log.rst
+++ b/docs/change_log.rst
@@ -56,6 +56,7 @@ The following are only available within Stem's `git repository
  * Support for ed25519 descriptor fields (:spec:`5a79d67`)
  * Server descriptor validation fails with 'extra-info-digest line had an invalid value' from additions in proposal 228 (:trac:`16227`)
  * :class:`~stem.descriptor.server_descriptor.BridgeDescriptor` now has 'ntor_onion_key' like its unsanitized counterparts
+  * Replaced the :class:`~stem.descriptor.microdescriptor.Microdescriptor` identifier and identifier_type attributes with an identifiers hash since it can now appear multiple times (:spec:`09ff9e2`)

 * **Website**


--- a/stem/descriptor/microdescriptor.py
+++ b/stem/descriptor/microdescriptor.py
@@ -73,6 +73,7 @@ from stem.descriptor import (
  _get_descriptor_components,
  _read_until_keywords,
  _value,
+  _values,
  _parse_simple_line,
  _parse_key_block,
 )
@@ -159,14 +160,24 @@ def _parse_file(descriptor_file, validate = False, **kwargs):


 def _parse_id_line(descriptor, entries):
-  value = _value('id', entries)
-  value_comp = value.split()
+  identities = {}

-  if len(value_comp) >= 2:
-    descriptor.identifier_type = value_comp[0]
-    descriptor.identifier = value_comp[1]
+  for entry in _values('id', entries):
+    entry_comp = entry.split()
+
+    if len(entry_comp) >= 2:
+      key_type, key_value = entry_comp[0], entry_comp[1]
+
+      if key_type in identities:
+        raise ValueError("There can only be one 'id' line per a key type, but '%s' appeared multiple times" % key_type)
+
+      descriptor.identifier_type = key_type
+      descriptor.identifier = key_value
+      identities[key_type] = key_value
    else:
-    raise ValueError("'id' lines should contain both the key type and digest: id %s" % value)
+      raise ValueError("'id' lines should contain both the key type and digest: id %s" % entry)
+
+  descriptor.identifiers = identities


 _parse_digest = lambda descriptor, entries: setattr(descriptor, 'digest', hashlib.sha256(descriptor.get_bytes()).hexdigest().upper())
@@ -192,13 +203,23 @@ class Microdescriptor(Descriptor):
  :var list family: **\*** nicknames or fingerprints of declared family
  :var stem.exit_policy.MicroExitPolicy exit_policy: **\*** relay's exit policy
  :var stem.exit_policy.MicroExitPolicy exit_policy_v6: **\*** exit policy for IPv6
-  :var str identifier_type: identity digest key type
-  :var str identifier: base64 encoded identity digest, this is only used for collision prevention (:trac:`11743`)
+  :var hash identifiers: mapping of key types (like rsa1024 or ed25519) to
+    their base64 encoded identity, this is only used for collision prevention
+    (:trac:`11743`)
+
+  :var str identifier: base64 encoded identity digest (**deprecated**, use
+    identifiers instead)
+  :var str identifier_type: identity digest key type (**deprecated**, use
+    identifiers instead)

  **\*** attribute is required when we're parsed with validation

  .. versionchanged:: 1.1.0
     Added the identifier and identifier_type attributes.
+
+  .. versionchanged:: 1.5.0
+     Added the identifiers attribute, and deprecated identifier and
+     identifier_type since the field can now appear multiple times.
  """

  ATTRIBUTES = {
@@ -208,8 +229,9 @@ class Microdescriptor(Descriptor):
    'family': ([], _parse_family_line),
    'exit_policy': (stem.exit_policy.MicroExitPolicy('reject 1-65535'), _parse_p_line),
    'exit_policy_v6': (None, _parse_p6_line),
-    'identifier_type': (None, _parse_id_line),
-    'identifier': (None, _parse_id_line),
+    'identifier_type': (None, _parse_id_line),  # deprecated in favor of identifiers
+    'identifier': (None, _parse_id_line),  # deprecated in favor of identifiers
+    'identifiers': ({}, _parse_id_line),
    'digest': (None, _parse_digest),
  }


--- a/stem/descriptor/server_descriptor.py
+++ b/stem/descriptor/server_descriptor.py
@@ -420,7 +420,7 @@ class ServerDescriptor(Descriptor):

  :var str address: **\*** IPv4 address of the relay
  :var int or_port: **\*** port used for relaying
-  :var int socks_port: **\*** port used as client (deprecated, always **None**)
+  :var int socks_port: **\*** port used as client (**deprecated**, always **None**)
  :var int dir_port: **\*** port used for descriptor mirroring

  :var bytes platform: line with operating system and tor version
@@ -442,13 +442,13 @@ class ServerDescriptor(Descriptor):
  :var bool allow_single_hop_exits: **\*** flag if single hop exiting is allowed
  :var bool extra_info_cache: **\*** flag if a mirror for extra-info documents
  :var str extra_info_digest: upper-case hex encoded digest of our extra-info document
-  :var bool eventdns: flag for evdns backend (deprecated, always unset)
+  :var bool eventdns: flag for evdns backend (**deprecated**, always unset)
  :var str ntor_onion_key: base64 key used to encrypt EXTEND in the ntor protocol
  :var list or_addresses: **\*** alternative for our address/or_port
    attributes, each entry is a tuple of the form (address (**str**), port
    (**int**), is_ipv6 (**bool**))

-  Deprecated, moved to extra-info descriptor...
+  **Deprecated**, moved to extra-info descriptor...

  :var datetime read_history_end: end of the sampling interval
  :var int read_history_interval: seconds per interval

--- a/test/unit/descriptor/microdescriptor.py
+++ b/test/unit/descriptor/microdescriptor.py
@@ -95,6 +95,7 @@ class TestMicrodescriptor(unittest.TestCase):
    self.assertEqual([], desc.family)
    self.assertEqual(stem.exit_policy.MicroExitPolicy('reject 1-65535'), desc.exit_policy)
    self.assertEqual(None, desc.exit_policy_v6)
+    self.assertEqual({}, desc.identifiers)
    self.assertEqual(None, desc.identifier_type)
    self.assertEqual(None, desc.identifier)
    self.assertEqual([], desc.get_unrecognized_lines())
@@ -170,5 +171,32 @@ class TestMicrodescriptor(unittest.TestCase):
    """

    desc = get_microdescriptor({'id': 'rsa1024 Cd47okjCHD83YGzThGBDptXs9Z4'})
+    self.assertEqual({'rsa1024': 'Cd47okjCHD83YGzThGBDptXs9Z4'}, desc.identifiers)
    self.assertEqual('rsa1024', desc.identifier_type)
    self.assertEqual('Cd47okjCHD83YGzThGBDptXs9Z4', desc.identifier)
+
+    # check when there's multiple key types
+
+    desc_text = b'\n'.join((get_microdescriptor(content = True),
+                            b'id rsa1024 Cd47okjCHD83YGzThGBDptXs9Z4',
+                            b'id ed25519 50f6ddbecdc848dcc6b818b14d1'))
+
+    desc = Microdescriptor(desc_text, validate = True)
+    self.assertEqual({'rsa1024': 'Cd47okjCHD83YGzThGBDptXs9Z4', 'ed25519': '50f6ddbecdc848dcc6b818b14d1'}, desc.identifiers)
+    self.assertEqual('ed25519', desc.identifier_type)
+    self.assertEqual('50f6ddbecdc848dcc6b818b14d1', desc.identifier)
+
+    # check when there's conflicting keys
+
+    desc_text = b'\n'.join((get_microdescriptor(content = True),
+                            b'id rsa1024 Cd47okjCHD83YGzThGBDptXs9Z4',
+                            b'id rsa1024 50f6ddbecdc848dcc6b818b14d1'))
+
+    desc = Microdescriptor(desc_text)
+    self.assertEqual({}, desc.identifiers)
+
+    try:
+      Microdescriptor(desc_text, validate = True)
+      self.fail('constructor validation should fail')
+    except ValueError as exc:
+      self.assertEqual("There can only be one 'id' line per a key type, but 'rsa1024' appeared multiple times", str(exc))