re_lookup shows the original source file (Xbox file/line sites) and the Mac build's original name

harbor/reverse.py

@@ -40,6 +40,8 @@ class Reverse:
         self.annotations = Path(config["annotations"]) if config.get("annotations") else None
         self.src = Path(config["src"]) if config.get("src") else None
         self.conventions_path = Path(config["conventions"]) if config.get("conventions") else None
+        self.source_files = Path(config["source_files"]) if config.get("source_files") else None
+        self.mac_names = Path(config["mac_names"]) if config.get("mac_names") else None
         self._cache = {}
 
     # ----------------------------------------------------------------- loading (cached by file mtime)
@@ -183,6 +185,32 @@ class Reverse:
         return {"summary": module.describe(result), "registers_in": result["registers_in"],
                 "ret_pops": result["ret_pops"]}
 
+    def _keyed_csv(self, key, path, column):
+        def load(p):
+            out = {}
+            if p:
+                for row in csv.DictReader(p.read_text(encoding="utf-8").splitlines()):
+                    try:
+                        out.setdefault(int(row[column], 16), row)
+                    except (KeyError, ValueError):
+                        pass
+            return out
+        return self._load(key, path, load)
+
+    def origins(self, addr):
+        """What the other builds say about a function: its original source file (Xbox file/line sites) and its
+        original name with parameter types (the unstripped Mac build). Evidence, not truth."""
+        out = {}
+        src = self._keyed_csv("source_files", self.source_files, "pc_address").get(addr)
+        if src:
+            out["source_file"] = {k: src.get(k) for k in ("source_file", "lines", "confidence", "xbox_address",
+                                                         "evidence") if src.get(k)}
+        mac = self._keyed_csv("mac_names", self.mac_names, "pc_address").get(addr)
+        if mac:
+            out["mac_name"] = {k: mac.get(k) for k in ("demangled", "mac_symbol", "confidence", "evidence")
+                               if mac.get(k)}
+        return out
+
     # ----------------------------------------------------------------- tools
     def read(self, addr, count):
         data = self._bytes()
@@ -289,6 +317,7 @@ class Reverse:
             conv = self.convention(func[0])
             if conv:
                 out["convention_from_bytes"] = conv
+            out.update(self.origins(func[0]))
         for table, key in (("names.csv", "names_row"), ("functions.csv", "functions_row"), ("globals.csv", "globals_row")):
             row = self._csv(table).get(addr)
             if row: