Add Ghidra function import script (#909)

* Add draft for Ghidra function import script

* feature: Basic PDB analysis [skip ci]

This is a draft with a lot of open questions left. Please do not merge

* Refactor: Introduce submodules and reload remedy

* refactor types and make them Python 3.9 compatible

* run black

* WIP: save progress

* fix types and small type safety violations

* fix another Python 3.9 syntax incompatibility

* Implement struct imports [skip ci]

- This code is still in dire need of refactoring and tests
- There are only single-digit issues left, and 2600 functions can be imported
- The biggest remaining error is mismatched stacks

* Refactor, implement enums, fix lots of bugs

* fix Python 3.9 issue

* refactor: address review comments

Not sure why VS Code suddenly decides to remove some empty spaces, but they don't make sense anyway

* add unit tests for new type parsers, fix linter issue

* refactor: db access from pdb_extraction.py

* Fix stack layout offset error

* fix: Undo incorrect reference change

* Fix CI issue

* Improve READMEs (fix typos, add information)

---------

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

tools/ghidra_scripts/lego_util/ghidra_helper.py

@@ -0,0 +1,100 @@
+"""A collection of helper functions for the interaction with Ghidra."""
+
+import logging
+
+from lego_util.exceptions import (
+    ClassOrNamespaceNotFoundInGhidraError,
+    TypeNotFoundInGhidraError,
+    MultipleTypesFoundInGhidraError,
+)
+
+# Disable spurious warnings in vscode / pylance
+# pyright: reportMissingModuleSource=false
+
+from ghidra.program.model.data import PointerDataType
+from ghidra.program.model.data import DataTypeConflictHandler
+from ghidra.program.flatapi import FlatProgramAPI
+from ghidra.program.model.data import DataType
+from ghidra.program.model.symbol import Namespace
+
+logger = logging.getLogger(__name__)
+
+
+def get_ghidra_type(api: FlatProgramAPI, type_name: str):
+    """
+    Searches for the type named `typeName` in Ghidra.
+
+    Raises:
+    - NotFoundInGhidraError
+    - MultipleTypesFoundInGhidraError
+    """
+    result = api.getDataTypes(type_name)
+    if len(result) == 0:
+        raise TypeNotFoundInGhidraError(type_name)
+    if len(result) == 1:
+        return result[0]
+
+    raise MultipleTypesFoundInGhidraError(type_name, result)
+
+
+def add_pointer_type(api: FlatProgramAPI, pointee: DataType) -> DataType:
+    new_data_type = PointerDataType(pointee)
+    new_data_type.setCategoryPath(pointee.getCategoryPath())
+    result_data_type = (
+        api.getCurrentProgram()
+        .getDataTypeManager()
+        .addDataType(new_data_type, DataTypeConflictHandler.KEEP_HANDLER)
+    )
+    if result_data_type is not new_data_type:
+        logger.debug(
+            "New pointer replaced by existing one. Fresh pointer: %s (class: %s)",
+            result_data_type,
+            result_data_type.__class__,
+        )
+    return result_data_type
+
+
+def get_ghidra_namespace(
+    api: FlatProgramAPI, namespace_hierachy: list[str]
+) -> Namespace:
+    namespace = api.getCurrentProgram().getGlobalNamespace()
+    for part in namespace_hierachy:
+        namespace = api.getNamespace(namespace, part)
+        if namespace is None:
+            raise ClassOrNamespaceNotFoundInGhidraError(namespace_hierachy)
+    return namespace
+
+
+def create_ghidra_namespace(
+    api: FlatProgramAPI, namespace_hierachy: list[str]
+) -> Namespace:
+    namespace = api.getCurrentProgram().getGlobalNamespace()
+    for part in namespace_hierachy:
+        namespace = api.getNamespace(namespace, part)
+        if namespace is None:
+            namespace = api.createNamespace(namespace, part)
+    return namespace
+
+
+def sanitize_name(name: str) -> str:
+    """
+    Takes a full class or function name and replaces characters not accepted by Ghidra.
+    Applies mostly to templates and names like `vbase destructor`.
+    """
+    new_class_name = (
+        name.replace("<", "[")
+        .replace(">", "]")
+        .replace("*", "#")
+        .replace(" ", "_")
+        .replace("`", "'")
+    )
+    if "<" in name:
+        new_class_name = "_template_" + new_class_name
+
+    if new_class_name != name:
+        logger.warning(
+            "Class or function name contains characters forbidden by Ghidra, changing from '%s' to '%s'",
+            name,
+            new_class_name,
+        )
+    return new_class_name