BETA10 import: Fix FUN_... collisions (#1105)

* Improve BETA10 name import, refactor

* Run formatter

---------

Co-authored-by: jonschz <jonschz@users.noreply.github.com>

tools/ghidra_scripts/import_functions_and_types_from_pdb.py

@@ -21,9 +21,7 @@
 # Disable spurious warnings in vscode / pylance
 # pyright: reportMissingModuleSource=false
 
-from enum import Enum
 import importlib
-from dataclasses import dataclass, field
 import logging.handlers
 import sys
 import logging
@@ -53,43 +51,8 @@ def reload_module(module: str):
 
 
 reload_module("lego_util.statistics")
-from lego_util.statistics import Statistics
-
-
-@dataclass
-class Globals:
-    verbose: bool
-    loglevel: int
-    running_from_ghidra: bool = False
-    # statistics
-    statistics: Statistics = field(default_factory=Statistics)
-
-
-class SupportedModules(Enum):
-    LEGO1 = 1
-    BETA10 = 2
-
-    def orig_filename(self):
-        if self == self.LEGO1:
-            return "LEGO1.DLL"
-        return "BETA10.DLL"
-
-    def recomp_filename_without_extension(self):
-        # in case we want to support more functions
-        return "LEGO1"
-
-    def build_dir_name(self):
-        if self == self.BETA10:
-            return "build_debug"
-        return "build"
-
-
-# hard-coded settings that we don't want to prompt in Ghidra every time
-GLOBALS = Globals(
-    verbose=False,
-    # loglevel=logging.INFO,
-    loglevel=logging.DEBUG,
-)
+reload_module("lego_util.globals")
+from lego_util.globals import GLOBALS, SupportedModules
 
 
 def setup_logging():
@@ -232,24 +195,24 @@ def main():
         origfile_name = getProgramFile().getName()
 
         if origfile_name == "LEGO1.DLL":
-            module = SupportedModules.LEGO1
+            GLOBALS.module = SupportedModules.LEGO1
         elif origfile_name in ["LEGO1D.DLL", "BETA10.DLL"]:
-            module = SupportedModules.BETA10
+            GLOBALS.module = SupportedModules.BETA10
         else:
             raise Lego1Exception(
                 f"Unsupported file name in import script: {origfile_name}"
             )
-    else:
-        module = SupportedModules.LEGO1
 
-    logger.info("Importing file: %s", module.orig_filename())
+    logger.info("Importing file: %s", GLOBALS.module.orig_filename())
 
     repo_root = get_repository_root()
-    origfile_path = repo_root.joinpath("legobin").joinpath(module.orig_filename())
-    build_directory = repo_root.joinpath(module.build_dir_name())
-    recompiledfile_name = f"{module.recomp_filename_without_extension()}.DLL"
+    origfile_path = repo_root.joinpath("legobin").joinpath(
+        GLOBALS.module.orig_filename()
+    )
+    build_directory = repo_root.joinpath(GLOBALS.module.build_dir_name())
+    recompiledfile_name = f"{GLOBALS.module.recomp_filename_without_extension()}.DLL"
     recompiledfile_path = build_directory.joinpath(recompiledfile_name)
-    pdbfile_name = f"{module.recomp_filename_without_extension()}.PDB"
+    pdbfile_name = f"{GLOBALS.module.recomp_filename_without_extension()}.PDB"
     pdbfile_path = build_directory.joinpath(pdbfile_name)
 
     if not GLOBALS.verbose:

tools/ghidra_scripts/lego_util/ghidra_helper.py

@@ -8,6 +8,7 @@ from lego_util.exceptions import (
     TypeNotFoundInGhidraError,
     MultipleTypesFoundInGhidraError,
 )
+from lego_util.globals import GLOBALS, SupportedModules
 
 # Disable spurious warnings in vscode / pylance
 # pyright: reportMissingModuleSource=false
@@ -105,6 +106,12 @@ def sanitize_name(name: str) -> str:
         .replace("`", "'")
     )
 
+    # Importing function names like `FUN_10001234` into BETA10 can be confusing
+    # because Ghidra's auto-generated functions look exactly the same.
+    # Therefore, such function names are replaced by `LEGO_10001234` in the BETA10 import.
+    if GLOBALS.module == SupportedModules.BETA10:
+        new_name = re.sub(r"FUN_([0-9a-f]{8})", r"LEGO1_\1", new_name)
+
     if "<" in name:
         new_name = "_template_" + new_name
 

tools/ghidra_scripts/lego_util/globals.py

@@ -0,0 +1,42 @@
+import logging
+from enum import Enum
+from dataclasses import dataclass, field
+from lego_util.statistics import Statistics
+
+
+class SupportedModules(Enum):
+    LEGO1 = 1
+    BETA10 = 2
+
+    def orig_filename(self):
+        if self == self.LEGO1:
+            return "LEGO1.DLL"
+        return "BETA10.DLL"
+
+    def recomp_filename_without_extension(self):
+        # in case we want to support more functions
+        return "LEGO1"
+
+    def build_dir_name(self):
+        if self == self.BETA10:
+            return "build_debug"
+        return "build"
+
+
+@dataclass
+class Globals:
+    verbose: bool
+    loglevel: int
+    module: SupportedModules
+    running_from_ghidra: bool = False
+    # statistics
+    statistics: Statistics = field(default_factory=Statistics)
+
+
+# hard-coded settings that we don't want to prompt in Ghidra every time
+GLOBALS = Globals(
+    verbose=False,
+    # loglevel=logging.INFO,
+    loglevel=logging.DEBUG,
+    module=SupportedModules.LEGO1,  # this default value will be used when run outside of Ghidra
+)