feat(compiler/runtime): use hash for generated class as id and refactor user type id encoding

.gitignore

@@ -28,6 +28,7 @@ java/**/dependency-reduced-pom.xml
 **/*.html
 **/target
 .idea/
+generated/
 java/dependency-reduced-pom.xml
 java/foo
 scala/.bsp
@@ -38,55 +39,10 @@ bazel-fory/**
 **/generated
 java/**/generated
 
-# IDL compiler integration test outputs
-integration_tests/idl_tests/cpp/generated/
-integration_tests/idl_tests/go/addressbook*.go
-integration_tests/idl_tests/go/any_example/
-integration_tests/idl_tests/go/any_example_pb/
-integration_tests/idl_tests/go/complex_fbs/
-integration_tests/idl_tests/go/complex_pb/
-integration_tests/idl_tests/go/monster/
-integration_tests/idl_tests/go/optional_types/
-integration_tests/idl_tests/go/tree/
-integration_tests/idl_tests/go/graph/
-integration_tests/idl_tests/go/collection/
-integration_tests/idl_tests/java/src/main/
-integration_tests/idl_tests/go/root/
-integration_tests/idl_tests/go/addressbook/
-integration_tests/idl_tests/java/src/main/java/collection/
-integration_tests/idl_tests/java/src/main/java/addressbook/
-integration_tests/idl_tests/java/src/main/java/complex_fbs/
-integration_tests/idl_tests/java/src/main/java/complex_pb/
-integration_tests/idl_tests/java/src/main/java/monster/
-integration_tests/idl_tests/java/src/main/java/optional_types/
-integration_tests/idl_tests/java/src/main/java/tree/
-integration_tests/idl_tests/java/src/main/java/graph/
-integration_tests/idl_tests/java/src/main/java/root/
-integration_tests/idl_tests/python/src/addressbook.py
-integration_tests/idl_tests/python/src/any_example.py
-integration_tests/idl_tests/python/src/any_example_pb.py
-integration_tests/idl_tests/python/src/complex_fbs.py
-integration_tests/idl_tests/python/src/complex_pb.py
-integration_tests/idl_tests/python/src/monster.py
-integration_tests/idl_tests/python/src/optional_types.py
-integration_tests/idl_tests/python/src/collection.py
-integration_tests/idl_tests/python/src/tree.py
-integration_tests/idl_tests/python/src/graph.py
-integration_tests/idl_tests/python/src/root.py
-integration_tests/idl_tests/rust/src/addressbook.rs
-integration_tests/idl_tests/rust/src/any_example.rs
-integration_tests/idl_tests/rust/src/any_example_pb.rs
-integration_tests/idl_tests/rust/src/complex_fbs.rs
-integration_tests/idl_tests/rust/src/complex_pb.rs
-integration_tests/idl_tests/rust/src/monster.rs
-integration_tests/idl_tests/rust/src/optional_types.rs
-integration_tests/idl_tests/rust/src/collection.rs
-integration_tests/idl_tests/rust/src/tree.rs
-integration_tests/idl_tests/rust/src/graph.rs
-integration_tests/idl_tests/rust/src/root.rs
 javascript/**/dist/
 javascript/**/node_modules/
 javascript/**/build
+javascript/junit.xml
 MODULE.bazel.lock
 .DS_Store
 **/.DS_Store

compiler/README.md

@@ -229,8 +229,12 @@ FDL uses plain option keys without a `(fory)` prefix:
 ```fdl
 option use_record_for_java_message = true;
 option polymorphism = true;
+option enable_auto_type_id = true;
 ```
 
+`enable_auto_type_id` defaults to `true`. Set it to `false` to keep namespace-based registration
+for types that omit explicit IDs.
+
 **Message/Enum options:**
 
 ```fdl

compiler/extension/fory_options.proto

@@ -28,6 +28,7 @@
 //
 // File-level options:
 //   option (fory).use_record_for_java_message = true;
+//   option (fory).enable_auto_type_id = true;
 //
 // Message options:
 //   message MyMessage {
@@ -67,6 +68,12 @@ message ForyFileOptions {
   // When true, type metadata is included in serialization for polymorphic
   // dispatch. Default: false
   optional bool polymorphism = 2;
+
+  // Control auto-generation of type IDs when none are specified.
+  // When true (default in the compiler), IDs are generated from the package
+  // namespace and type name. When false, types without explicit IDs are
+  // registered by namespace and name.
+  optional bool enable_auto_type_id = 3;
 }
 
 extend google.protobuf.FileOptions { optional ForyFileOptions fory = 50001; }

compiler/fory_compiler/cli.py

@@ -153,6 +153,7 @@ def resolve_imports(
     # Create merged schema with imported types first (so they can be referenced)
     merged_schema = Schema(
         package=schema.package,
+        package_alias=schema.package_alias,
         imports=schema.imports,
         enums=imported_enums + schema.enums,
         messages=imported_messages + schema.messages,
@@ -181,8 +182,34 @@ def go_package_info(schema: Schema) -> Tuple[Optional[str], str]:
     return None, ""
 
 
+def _find_go_module_root(base_go_out: Path) -> Optional[Path]:
+    base_go_out = base_go_out.resolve()
+    for candidate in (base_go_out, *base_go_out.parents):
+        if (candidate / "go.mod").is_file():
+            return candidate
+    return None
+
+
+def _read_go_module_path(go_module_root: Path) -> Optional[str]:
+    module_file = go_module_root / "go.mod"
+    if not module_file.is_file():
+        return None
+    try:
+        with module_file.open("r", encoding="utf-8") as handle:
+            for line in handle:
+                stripped = line.strip()
+                if stripped.startswith("module "):
+                    return stripped.split(None, 1)[1].strip()
+    except OSError:
+        return None
+    return None
+
+
 def resolve_go_module_root(base_go_out: Path, schema: Schema) -> Path:
     """Infer the Go module root for output layout."""
+    module_root = _find_go_module_root(base_go_out)
+    if module_root:
+        return module_root
     _, package_name = go_package_info(schema)
     if package_name and base_go_out.name == package_name:
         return base_go_out.parent
@@ -195,6 +222,14 @@ def resolve_go_output_dir(base_go_out: Path, schema: Schema) -> Path:
     if package_name and base_go_out.name == package_name:
         return base_go_out
     if import_path:
+        module_path = _read_go_module_path(base_go_out)
+        if module_path:
+            module_path = module_path.rstrip("/")
+            if import_path.startswith(module_path):
+                rel_path = import_path[len(module_path) :].lstrip("/")
+                if rel_path:
+                    return base_go_out / rel_path
+                return base_go_out
         last_segment = import_path.rstrip("/").split("/")[-1]
         if package_name and last_segment == package_name:
             return base_go_out / package_name
@@ -501,7 +536,13 @@ def compile_file_recursive(
     effective_outputs = lang_output_dirs
     if "go" in lang_output_dirs:
         go_root = go_module_root or lang_output_dirs["go"]
-        go_out = resolve_go_output_dir(go_root, schema)
+        if (
+            schema.get_option("go_package") is None
+            and lang_output_dirs["go"] != go_root
+        ):
+            go_out = lang_output_dirs["go"]
+        else:
+            go_out = resolve_go_output_dir(go_root, schema)
         if go_out != lang_output_dirs["go"]:
             effective_outputs = dict(lang_output_dirs)
             effective_outputs["go"] = go_out

compiler/fory_compiler/frontend/fbs/translator.py

@@ -77,6 +77,7 @@ def _location(self, line: int, column: int) -> SourceLocation:
     def translate(self) -> Schema:
         return Schema(
             package=self.schema.namespace,
+            package_alias=None,
             imports=[Import(path=inc) for inc in self.schema.includes],
             enums=[self._translate_enum(e) for e in self.schema.enums],
             unions=[self._translate_union(u) for u in self.schema.unions],

compiler/fory_compiler/frontend/fdl/parser.py

@@ -18,7 +18,7 @@
 """Recursive descent parser for FDL."""
 
 import warnings
-from typing import List, Set
+from typing import List, Set, Optional
 
 from fory_compiler.ir.ast import (
     Schema,
@@ -47,6 +47,7 @@
     "deprecated",
     "use_record_for_java_message",
     "polymorphism",
+    "enable_auto_type_id",
     "go_nested_type_style",
 }
 
@@ -92,6 +93,7 @@
 
 KNOWN_MESSAGE_OPTIONS: Set[str] = {
     "id",
+    "alias",
     "evolving",
     "use_record_for_java",
     "deprecated",
@@ -100,6 +102,7 @@
 
 KNOWN_UNION_OPTIONS: Set[str] = {
     "id",
+    "alias",
     "deprecated",
 }
 
@@ -187,6 +190,7 @@ def error(self, message: str) -> ParseError:
     def parse(self) -> Schema:
         """Parse the entire input and return a Schema."""
         package = None
+        package_alias = None
         imports = []
         enums = []
         messages = []
@@ -197,7 +201,7 @@ def parse(self) -> Schema:
             if self.check(TokenType.PACKAGE):
                 if package is not None:
                     raise self.error("Duplicate package declaration")
-                package = self.parse_package()
+                package, package_alias = self.parse_package()
             elif self.check(TokenType.IMPORT):
                 imports.append(self.parse_import())
             elif self.check(TokenType.OPTION):
@@ -215,6 +219,7 @@ def parse(self) -> Schema:
 
         return Schema(
             package=package,
+            package_alias=package_alias,
             imports=imports,
             enums=enums,
             messages=messages,
@@ -233,8 +238,8 @@ def make_location(self, token: Token) -> SourceLocation:
             source_format=self.source_format,
         )
 
-    def parse_package(self) -> str:
-        """Parse a package declaration: package foo.bar;"""
+    def parse_package(self) -> tuple[str, Optional[str]]:
+        """Parse a package declaration: package foo.bar [alias baz];"""
         self.consume(TokenType.PACKAGE)
 
         # Package name can be dotted: foo.bar.baz
@@ -245,8 +250,23 @@ def parse_package(self) -> str:
                 self.consume(TokenType.IDENT, "Expected identifier after '.'").value
             )
 
-        self.consume(TokenType.SEMI, "Expected ';' after package name")
-        return ".".join(parts)
+        alias = None
+        if self.check(TokenType.IDENT) and self.current().value == "alias":
+            self.advance()  # consume alias keyword
+            alias_parts = [
+                self.consume(TokenType.IDENT, "Expected identifier after 'alias'").value
+            ]
+            while self.check(TokenType.DOT):
+                self.advance()
+                alias_parts.append(
+                    self.consume(
+                        TokenType.IDENT, "Expected identifier after '.' in alias"
+                    ).value
+                )
+            alias = ".".join(alias_parts)
+
+        self.consume(TokenType.SEMI, "Expected ';' after package declaration")
+        return ".".join(parts), alias
 
     def parse_option_value(self):
         """Parse an option value (string, bool, int, or identifier)."""

compiler/fory_compiler/frontend/proto/translator.py

@@ -97,6 +97,7 @@ def _location(self, line: int, column: int) -> SourceLocation:
     def translate(self) -> Schema:
         return Schema(
             package=self.proto_schema.package,
+            package_alias=None,
             imports=self._translate_imports(),
             enums=[self._translate_enum(e) for e in self.proto_schema.enums],
             messages=[self._translate_message(m) for m in self.proto_schema.messages],

compiler/fory_compiler/generators/base.py

@@ -181,6 +181,21 @@ def strip_enum_prefix(self, enum_name: str, value_name: str) -> str:
 
         return remainder
 
+    def format_type_id_comment(self, type_def, comment_prefix: str) -> Optional[str]:
+        """Format a type id comment for a message/union."""
+        type_id = getattr(type_def, "type_id", None)
+        if type_id is None:
+            return None
+        if getattr(type_def, "id_generated", False):
+            source = getattr(type_def, "id_source", None) or "unknown"
+            return f"{comment_prefix} Type ID {type_id} is generated from {source}"
+        return f"{comment_prefix} Type ID {type_id} is specified manually."
+
+    def should_register_by_id(self, type_def) -> bool:
+        """Return True if a type should be registered by numeric ID."""
+        type_id = getattr(type_def, "type_id", None)
+        return type_id is not None
+
     def get_license_header(self, comment_prefix: str = "//") -> str:
         """Get the Apache license header."""
         lines = [

compiler/fory_compiler/generators/cpp.py

@@ -880,6 +880,9 @@ def generate_message_definition(
         lineage = parent_stack + [message]
         body_indent = f"{indent}  "
         field_indent = f"{indent}    "
+        comment = self.format_type_id_comment(message, f"{indent}//")
+        if comment:
+            lines.append(comment)
         lines.append(f"{indent}class {class_name} final {{")
         lines.append(f"{body_indent}public:")
         if message.fields:
@@ -982,6 +985,9 @@ def generate_union_definition(
         ]
         variant_type = f"std::variant<{', '.join(case_types)}>"
 
+        comment = self.format_type_id_comment(union, f"{indent}//")
+        if comment:
+            lines.append(comment)
         lines.append(f"{indent}class {class_name} final {{")
         lines.append(f"{body_indent}public:")
         lines.append(f"{body_indent}  enum class {case_enum} : uint32_t {{")
@@ -1234,7 +1240,7 @@ def generate_union_serializer(
         lines.append("")
         lines.append("  static inline void write_type_info(WriteContext &ctx) {")
         lines.append(
-            f"    auto result = ctx.write_any_typeinfo(static_cast<uint32_t>(TypeId::TYPED_UNION), std::type_index(typeid({qualified_name})));"
+            f"    auto result = ctx.write_any_type_info(static_cast<uint32_t>(TypeId::TYPED_UNION), std::type_index(typeid({qualified_name})));"
         )
         lines.append("    if (FORY_PREDICT_FALSE(!result.ok())) {")
         lines.append("      ctx.set_error(std::move(result).error());")
@@ -1250,7 +1256,9 @@ def generate_union_serializer(
         lines.append("      return;")
         lines.append("    }")
         lines.append("    const TypeInfo *expected = type_info_res.value();")
-        lines.append("    const TypeInfo *remote = ctx.read_any_typeinfo(ctx.error());")
+        lines.append(
+            "    const TypeInfo *remote = ctx.read_any_type_info(ctx.error());"
+        )
         lines.append("    if (FORY_PREDICT_FALSE(ctx.has_error())) {")
         lines.append("      return;")
         lines.append("    }")
@@ -1391,7 +1399,7 @@ def generate_union_serializer(
         lines.append("        return default_value();")
         lines.append("      }")
         lines.append(
-            "      const TypeInfo *type_info = ctx.read_any_typeinfo(ctx.error());"
+            "      const TypeInfo *type_info = ctx.read_any_type_info(ctx.error());"
         )
         lines.append("      if (FORY_PREDICT_FALSE(ctx.has_error())) {")
         lines.append("        return default_value();")
@@ -1858,7 +1866,7 @@ def generate_enum_registration(
         code_name = self.get_qualified_type_name(enum.name, parent_stack)
         type_name = self.get_registration_type_name(enum.name, parent_stack)
 
-        if enum.type_id is not None:
+        if self.should_register_by_id(enum):
             lines.append(f"    fory.register_enum<{code_name}>({enum.type_id});")
         else:
             ns = self.package or "default"
@@ -1889,7 +1897,7 @@ def generate_message_registration(
             )
 
         # Register this message
-        if message.type_id is not None:
+        if self.should_register_by_id(message):
             lines.append(f"    fory.register_struct<{code_name}>({message.type_id});")
         else:
             ns = self.package or "default"
@@ -1904,7 +1912,7 @@ def generate_union_registration(
         code_name = self.get_qualified_type_name(union.name, parent_stack)
         type_name = self.get_registration_type_name(union.name, parent_stack)
 
-        if union.type_id is not None:
+        if self.should_register_by_id(union):
             lines.append(f"    fory.register_union<{code_name}>({union.type_id});")
         else:
             ns = self.package or "default"