Use int as basis for CvdumpTypeKey

reccmp/compare/ingest.py

@@ -29,6 +29,7 @@ def load_cvdump_types(cvdump_analysis: CvdumpAnalysis, types: CvdumpTypesParser)
     # TODO: Populate the universal type database here when this exists. (#106)
     # For now, just copy the keys into another CvdumpTypesParser so we can use its API.
     types.keys.update(cvdump_analysis.types.keys)
+    types.alerted_types = cvdump_analysis.types.alerted_types
 
 
 def load_cvdump(cvdump_analysis: CvdumpAnalysis, db: EntityDb, recomp_bin: PEImage):
@@ -111,6 +112,7 @@ def load_cvdump(cvdump_analysis: CvdumpAnalysis, db: EntityDb, recomp_bin: PEIma
 
                 # Set the cvdump type key so it can be referenced later.
                 if sym.node_type == EntityType.DATA and sym.data_type is not None:
+                    assert isinstance(sym.data_type.key, int)
                     batch.set_recomp(addr, data_type=sym.data_type.key)
 
 

reccmp/compare/mutate.py

@@ -8,6 +8,7 @@
     get_function_arg_string,
 )
 from reccmp.cvdump import CvdumpTypesParser
+from reccmp.cvdump.types import CvdumpTypeKey
 from reccmp.types import EntityType
 from .db import EntityDb
 from .queries import get_overloaded_functions, get_named_thunks
@@ -27,7 +28,7 @@ def match_array_elements(db: EntityDb, types: CvdumpTypesParser):
     batch = db.batch()
 
     @cache
-    def get_type_size(type_key: str) -> int:
+    def get_type_size(type_key: CvdumpTypeKey) -> int:
         type_ = types.get(type_key)
         assert type_.size is not None
         return type_.size
@@ -57,15 +58,16 @@ def _add_match_in_array(
 
     for match in db.get_matches_by_type(EntityType.DATA):
         # TODO: The type information we need is in multiple places. (See #106)
-        type_key = match.get("data_type")
-        if type_key is None:
+        type_key_raw = match.get("data_type")
+        if type_key_raw is None:
             continue
 
-        if not type_key.startswith("0x"):
+        type_key = CvdumpTypeKey(type_key_raw)
+        if type_key.is_scalar():
             # scalar type, so clearly not an array
             continue
 
-        type_dict = types.keys.get(type_key.lower())
+        type_dict = types.keys.get(type_key)
         if type_dict is None:
             continue
 
@@ -76,7 +78,7 @@ def _add_match_in_array(
         if array_type_key is None:
             continue
 
-        data_type = types.get(type_key.lower())
+        data_type = types.get(type_key)
 
         # Check whether another orig variable appears before the end of the array in recomp.
         # If this happens we can still add all the recomp offsets, but do not attach the orig address

reccmp/cvdump/analysis.py

@@ -6,7 +6,12 @@
 from .demangler import demangle_vtable
 from .parser import CvdumpParser, LineValue, NodeKey
 from .symbols import SymbolsEntry
-from .types import CvdumpKeyError, CvdumpIntegrityError, CvdumpTypesParser, TypeInfo
+from .types import (
+    CvdumpKeyError,
+    CvdumpIntegrityError,
+    CvdumpTypesParser,
+    TypeInfo,
+)
 
 
 @dataclass

reccmp/cvdump/parser.py

@@ -3,6 +3,7 @@
 from typing import NamedTuple
 from .types import CvdumpTypesParser
 from .symbols import CvdumpSymbolsParser
+from .cvinfo import CvdumpTypeKey
 
 
 # e.g. `     27 00034EC0     28 00034EE2     29 00034EE7     30 00034EF4`
@@ -76,7 +77,7 @@ class GdataEntry(NamedTuple):
 
     section: int
     offset: int
-    type: str
+    type: CvdumpTypeKey
     name: str
     is_global: bool
 
@@ -165,7 +166,7 @@ def _globals_section(self, line: str):
                 GdataEntry(
                     section=int(match.group("section"), 16),
                     offset=int(match.group("offset"), 16),
-                    type=match.group("type"),
+                    type=CvdumpTypeKey.from_str(match.group("type")),
                     name=match.group("name"),
                     is_global=match.group("global") == "G",
                 )

reccmp/cvdump/symbols.py

@@ -3,6 +3,7 @@
 import re
 from re import Match
 from typing import NamedTuple
+from reccmp.cvdump.types import CvdumpTypeKey
 
 
 logger = logging.getLogger(__name__)
@@ -12,7 +13,7 @@ class StackOrRegisterSymbol(NamedTuple):
     symbol_type: str
     location: str
     """Should always be set/converted to lowercase."""
-    data_type: str
+    data_type: CvdumpTypeKey
     name: str
 
 
@@ -21,7 +22,7 @@ class LdataEntry(NamedTuple):
 
     section: int
     offset: int
-    type: str
+    type: CvdumpTypeKey
     name: str
 
 
@@ -33,7 +34,7 @@ class SymbolsEntry:
     section: int
     offset: int
     size: int
-    func_type: str
+    func_type: CvdumpTypeKey
     name: str
     stack_symbols: list[StackOrRegisterSymbol] = field(default_factory=list)
     static_variables: list[LdataEntry] = field(default_factory=list)
@@ -165,7 +166,7 @@ def _parse_generic_case(self, line, line_match: Match[str]):
                 section=int(match.group("section"), 16),
                 offset=int(match.group("offset"), 16),
                 size=int(match.group("size"), 16),
-                func_type=match.group("func_type"),
+                func_type=CvdumpTypeKey.from_str(match.group("func_type")),
                 name=match.group("name"),
             )
             self.symbols.append(self.current_function)
@@ -182,7 +183,7 @@ def _parse_generic_case(self, line, line_match: Match[str]):
             new_symbol = StackOrRegisterSymbol(
                 symbol_type=symbol_type,
                 location=match.group("location").lower(),
-                data_type=match.group("data_type"),
+                data_type=CvdumpTypeKey.from_str(match.group("data_type")),
                 name=match.group("name"),
             )
             self.current_function.stack_symbols.append(new_symbol)
@@ -201,7 +202,7 @@ def _parse_generic_case(self, line, line_match: Match[str]):
                 new_var = LdataEntry(
                     section=int(match.group("section"), 16),
                     offset=int(match.group("offset"), 16),
-                    type=match.group("type"),
+                    type=CvdumpTypeKey.from_str(match.group("type")),
                     name=match.group("name"),
                 )