docs(tests): stop calling pull requests issues in test prose

tests/corekit/test_enum_lookup_reparent_930_unit.py

@@ -2,7 +2,7 @@
 """GitHub issue #930, the blocked half of #877's phase 2: re-parenting the last
 seven non-registry enumerations onto :class:`~pcapkit.corekit.enum.EnumLookup`.
 
-GitHub issue #921 re-parented 17 of the 24 non-registry enumerations and deliberately
+GitHub pull request #921 re-parented 17 of the 24 non-registry enumerations and deliberately
 left seven alone, because the files holding them were still open under other pull
 requests at the time: :mod:`pcapkit.const.ftp.command` (``CommandType``,
 ``ConformanceRequirement``) under #913, and :mod:`pcapkit.protocols.internet.esp`

tests/corekit/test_fields_numbers_port_option_no_mint_unit.py

@@ -232,7 +232,7 @@ def test_a_documented_span_still_mints_its_own_bounded_member(self) -> None:
                 self.assertEqual(resolved.port, 230)
 
     def test_an_out_of_width_port_still_raises_764s_rejection(self) -> None:
-        """The bounded fallback must not revert GitHub issue #764.
+        """The bounded fallback must not revert GitHub pull request #764.
 
         ``post_process`` is reachable directly -- not only through ``unpack``,
         which cannot even construct an out-of-range value from a 2-octet

tests/corekit/test_sentinel_exports_unit.py

@@ -234,7 +234,7 @@ def test_module_exports_the_object_and_not_the_type(self) -> 'None':
     def test_enum_exports_the_object_and_not_the_type(self) -> 'None':
         """``EnumLookup`` and ``EnumRegistry`` are not sentinels and stay.
 
-        ``EnumLookup`` in particular: GitHub issue #906 split it out as a public
+        ``EnumLookup`` in particular: GitHub pull request #906 split it out as a public
         base and
         :file:`docs/source/contributing/conventions/registry-protocol.rst` cites
         its ``get``, so dropping it while removing the sentinel type next to it

tests/foundation/registry/test_protocols.py

@@ -532,7 +532,7 @@ def test_top_level_link_internet_and_transport_protocol_wrappers(self) -> None:
                                       TransportProtocol.tcp | TransportProtocol.udp)
 
         # NOTE: a ``str`` transport is coerced to the member with that name --
-        # GitHub issue #815 ruling -- so a single string reaches the same
+        # GitHub pull request #815 ruling -- so a single string reaches the same
         # registry the equivalent member would.
         with mock.patch.object(registry.TCP, 'register') as tcp_register:
             with mock.patch.object(registry.UDP, 'register') as udp_register:

tests/project/test_conventions_doc_claims.py

@@ -802,7 +802,7 @@ def test_the_qualified_sentinel_targets_stay_qualified(self) -> 'None':
         """A later edit unqualifying one of these reintroduces #934 part B's miss.
 
         ``AbsentType``, ``NoValueType`` and ``ABSENT`` each resolve only against
-        the sentinels page's own module context (GitHub issue #936); written bare
+        the sentinels page's own module context (GitHub pull request #936); written bare
         anywhere on these pages, none of the three resolves at all. All four
         qualified references happen to live in
         :file:`sentinel-convention.rst`, but this scans every split page plus the

tests/protocols/schema/test_enum_schema_registry_unit.py

@@ -380,7 +380,7 @@ def test_a_repeated_code_in_the_declaration_list_stays_quiet(self) -> None:
         possible -- reaches the same key twice with ``cls`` on both sides,
         with no second call to ``__init_subclass__`` needed: this is a live
         path through normal syntax, not merely the direct-call case pinned
-        above. Not a contrived shape either: GitHub issue #721 shipped
+        above. Not a contrived shape either: GitHub pull request #721 shipped
         ``R1CounterParameter(Parameter, code=[R1_Counter, R1_COUNTER])``, and
         several :mod:`pcapkit.const` enums (e.g. ``reg.linktype.LinkType``,
         ``esp.cipher.Cipher``) declare real aliases that would make an

tests/protocols/test_dispatch_default_resolution_unit.py

@@ -5,7 +5,7 @@
 produces -- normally :class:`~pcapkit.protocols.misc.raw.Raw`. That resolution is
 deliberately **not** written back into the registry, because the registry is a
 class-level :class:`collections.defaultdict` and recording a miss in it is the
-defect GitHub issues #425/#428 fixed at this layer and #560 fixed at the schema
+defect GitHub issue #425 reported and pull request #428 fixed at this layer, and #560 fixed at the schema
 layer. The cost of not writing it back is that every unrecognised frame resolves
 the same descriptor again: 48 of the 52 resolutions an extraction of
 :file:`many_interfaces.pcapng` performs.

tests/test_docstring_contract.py

@@ -1,7 +1,7 @@
 # -*- coding: utf-8 -*-
 """Every ``Args:`` name, ``Raises:`` clause, and quoted RFC sentence, checked.
 
-GitHub issue #501 fixed one wrong exception name and forty stale ``Args:``
+GitHub pull request #501 fixed one wrong exception name and forty stale ``Args:``
 labels by hand. Issue #519 then showed the class was not exhausted -- more
 phantom ``Raises:`` clauses and more ``Args:`` entries naming parameters that do
 not exist -- because nothing in the suite *derives* the answer from the code.

tests/utilities/test_exceptions_excepthook.py

@@ -437,7 +437,7 @@ def exploding_delegate(etype, value, tb):
 class CrossInstanceReinstallTests(unittest.TestCase):
     """A fresh module instance must install *over* a stale instance's hook.
 
-    GitHub issue #983: CI caught this one. ``_install_excepthook``'s guard used
+    GitHub pull request #983: CI caught this one. ``_install_excepthook``'s guard used
     to skip installing whenever :data:`sys.excepthook` already carried the
     ``installed_by_pcapkit`` marker -- a plain ``True``, identical on *every*
     reloaded copy of :func:`~pcapkit.utilities.exceptions._excepthook`, so it