Mark ObjC declarations with _callable / _callable_class (#3228)

graphify/extractors/objc.py

@@ -12,6 +12,14 @@
 # `C++Bridge.h` and `Foo+.h` are left intact.
 _OBJC_STEM_PART = re.compile(r"[A-Za-z_][A-Za-z0-9_]*")
 
+# Declaration markers, matching what the generic engine puts on every other
+# language's definitions (#2438): `_callable` says "a real callable, not a
+# same-named data symbol", and `_callable_class` narrows that to a type, which is
+# callable only through a constructor. Passes that index declarations gate on
+# these, so a node without them is invisible to them.
+_CALLABLE = ("_callable",)
+_CALLABLE_CLASS = ("_callable", "_callable_class")
+
 
 def _objc_category_base_stem(stem: str) -> str:
     """Strip an ObjC category/extension suffix from a file stem (``Foo+Cat`` -> ``Foo``).
@@ -115,11 +123,14 @@ def extract_objc(path: Path) -> dict:
     # same (class, field) tombstones the entry (None) — drop, don't guess.
     objc_field_types: dict[str, dict[str, str | None]] = {}
 
-    def add_node(nid: str, label: str, line: int) -> None:
+    def add_node(nid: str, label: str, line: int, markers: tuple[str, ...] = ()) -> None:
         if nid not in seen_ids:
             seen_ids.add(nid)
-            nodes.append({"id": nid, "label": label, "file_type": "code",
-                          "source_file": str_path, "source_location": f"L{line}"})
+            node = {"id": nid, "label": label, "file_type": "code",
+                    "source_file": str_path, "source_location": f"L{line}"}
+            for marker in markers:
+                node[marker] = True
+            nodes.append(node)
 
     def add_edge(src: str, tgt: str, relation: str, line: int,
                  confidence: str = "EXTRACTED", weight: float = 1.0,
@@ -284,7 +295,7 @@ def walk(node, parent_nid: str | None = None) -> None:
             # produced fine when the members lived in `Foo.h` (#1556).
             cls_stem = _objc_category_base_stem(stem) if _objc_is_category(node) else stem
             cls_nid = _make_id(cls_stem, name)
-            add_node(cls_nid, name, line)
+            add_node(cls_nid, name, line, _CALLABLE_CLASS)
             add_edge(file_nid, cls_nid, "contains", line)
             # superclass is second identifier after ':'
             colon_seen = False
@@ -349,7 +360,7 @@ def walk(node, parent_nid: str | None = None) -> None:
             impl_stem = _objc_category_base_stem(stem) if _objc_is_category(node) else stem
             impl_nid = _make_id(impl_stem, name)
             if impl_nid not in seen_ids:
-                add_node(impl_nid, name, line)
+                add_node(impl_nid, name, line, _CALLABLE_CLASS)
                 add_edge(file_nid, impl_nid, "contains", line)
             for child in node.children:
                 if child.type == "instance_variables":
@@ -367,7 +378,9 @@ def walk(node, parent_nid: str | None = None) -> None:
                     break
             if name:
                 proto_nid = _make_id(stem, name)
-                add_node(proto_nid, f"<{name}>", line)
+                # A protocol is a type declaration like any other interface, and
+                # the engine marks a Java/C# interface the same way.
+                add_node(proto_nid, f"<{name}>", line, _CALLABLE_CLASS)
                 add_edge(file_nid, proto_nid, "contains", line)
                 # Adopted protocols: `@protocol Derived <Base, Other>`. These
                 # nest under a protocol_reference_list node (distinct from the
@@ -402,7 +415,7 @@ def walk(node, parent_nid: str | None = None) -> None:
             method_name = "".join(parts) if parts else None
             if method_name:
                 method_nid = _make_id(container, method_name)
-                add_node(method_nid, f"{prefix}{method_name}", line)
+                add_node(method_nid, f"{prefix}{method_name}", line, _CALLABLE)
                 add_edge(container, method_nid, "method", line)
                 if t == "method_definition":
                     method_bodies.append((method_nid, node, container))

tests/test_objc_callable_markers.py

@@ -0,0 +1,106 @@
+"""ObjC declarations carry the `_callable` / `_callable_class` markers.
+
+Every other extractor stamps its definitions with `_callable` — "a real callable,
+not a same-named data symbol" (#2438) — and narrows a type to `_callable_class`,
+callable only through a constructor (#2137). The generic engine does it for all of
+them in one place (`extractors/engine.py`, `callable_def_nids` /
+`callable_class_nids`).
+
+The ObjC extractor builds its nodes by hand and set neither, so an ObjC class was
+invisible to every pass that indexes declarations by those markers, and an ObjC
+method could never be told apart from a data symbol of the same name. This pins the
+markers onto the four node kinds ObjC produces, including the two that must stay
+unmarked.
+"""
+from __future__ import annotations
+
+from pathlib import Path
+
+from graphify.extract import extract
+
+GREETER_H = (
+    "@interface Greeter : NSObject\n"
+    "- (void)greet;\n"
+    "+ (instancetype)shared;\n"
+    "@end\n"
+)
+GREETER_M = (
+    "#import \"Greeter.h\"\n"
+    "@implementation Greeter\n"
+    "- (void)greet {}\n"
+    "@end\n"
+)
+
+
+def _extract(tmp_path: Path, files: dict[str, str]) -> dict:
+    paths = []
+    for name, body in files.items():
+        path = tmp_path / name
+        path.parent.mkdir(parents=True, exist_ok=True)
+        path.write_text(body, encoding="utf-8")
+        paths.append(path)
+    return extract(paths, cache_root=tmp_path / "graphify-out")
+
+
+def _node(result: dict, label: str) -> dict:
+    matches = [n for n in result["nodes"] if n.get("label") == label]
+    assert len(matches) == 1, [n.get("label") for n in result["nodes"]]
+    return matches[0]
+
+
+def test_a_class_is_marked_as_a_type(tmp_path: Path):
+    result = _extract(tmp_path, {"Greeter.h": GREETER_H})
+    greeter = _node(result, "Greeter")
+    assert greeter.get("_callable") is True
+    assert greeter.get("_callable_class") is True
+
+
+def test_an_implementation_only_class_is_marked_too(tmp_path: Path):
+    # A class whose `@interface` is not in this corpus is still a declaration.
+    result = _extract(tmp_path, {"Greeter.m": "@implementation Greeter\n- (void)greet {}\n@end\n"})
+    greeter = _node(result, "Greeter")
+    assert greeter.get("_callable") is True
+    assert greeter.get("_callable_class") is True
+
+
+def test_a_header_and_impl_pair_keeps_the_markers_after_folding(tmp_path: Path):
+    # `_merge_decl_def_classes` folds the `.h`/`.m` pair into one node; the markers
+    # have to be on whichever node survives.
+    result = _extract(tmp_path, {"Greeter.h": GREETER_H, "Greeter.m": GREETER_M})
+    greeter = _node(result, "Greeter")
+    assert greeter.get("_callable") is True
+    assert greeter.get("_callable_class") is True
+
+
+def test_a_method_is_callable_but_is_not_a_type(tmp_path: Path):
+    result = _extract(tmp_path, {"Greeter.h": GREETER_H})
+    for label in ("-greet", "+shared"):
+        method = _node(result, label)
+        assert method.get("_callable") is True, label
+        assert "_callable_class" not in method, label
+
+
+def test_a_protocol_is_marked_like_any_other_interface(tmp_path: Path):
+    # A Java or C# interface gets `_callable_class` from the generic engine, and a
+    # protocol is the same kind of declaration.
+    result = _extract(tmp_path, {"Greeting.h": "@protocol Greeting\n- (void)greet;\n@end\n"})
+    protocol = _node(result, "<Greeting>")
+    assert protocol.get("_callable") is True
+    assert protocol.get("_callable_class") is True
+
+
+def test_a_dangling_reference_is_not_marked(tmp_path: Path):
+    # `NSObject` is a stub minted for a name this corpus never declares, so it has
+    # no source file and no declaration behind it.
+    result = _extract(tmp_path, {"Greeter.h": GREETER_H})
+    stub = _node(result, "NSObject")
+    assert not stub.get("source_file")
+    assert "_callable" not in stub
+    assert "_callable_class" not in stub
+
+
+def test_the_file_node_is_not_marked(tmp_path: Path):
+    result = _extract(tmp_path, {"Greeter.h": GREETER_H})
+    file_node = _node(result, "Greeter.h")
+    assert "_callable" not in file_node
+    assert "_callable_class" not in file_node