Add a strict mode option. In the future we should refactor all options into a single Options dataclass.

Author: pavel-kirienko

pydsdl/__init__.py

@@ -11,7 +11,7 @@
 __version_info__ = tuple(map(int, __version__.split(".")[:3]))
 __license__ = "MIT"
 __author__ = "OpenCyphal"
-__copyright__ = "Copyright (c) 2018 OpenCyphal"
+__copyright__ = "Copyright (c) OpenCyphal development team"
 __email__ = "[email protected]"
 
 # Our unorthodox approach to dependency management requires us to apply certain workarounds.

pydsdl/_dsdl.py

@@ -119,6 +119,8 @@ def read(
         definition_visitors: Iterable["DefinitionVisitor"],
         print_output_handler: Callable[[int, str], None],
         allow_unregulated_fixed_port_id: bool,
+        *,
+        strict: bool = False,
     ) -> CompositeType:
         """
         Reads the data type definition and returns its high-level data type representation.
@@ -133,6 +135,7 @@ def read(
         :param definition_visitors:             Visitors to notify about discovered dependencies.
         :param print_output_handler:            Used for @print and for diagnostics: (line_number, text) -> None.
         :param allow_unregulated_fixed_port_id: Do not complain about fixed unregulated port IDs.
+        :param strict:                          Reject features that are not part of the Specification.
         :return: The data type representation.
         """
         raise NotImplementedError()

pydsdl/_dsdl_definition.py

@@ -235,6 +235,8 @@ def read(
         definition_visitors: Iterable[DefinitionVisitor],
         print_output_handler: Callable[[int, str], None],
         allow_unregulated_fixed_port_id: bool,
+        *,
+        strict: bool = False,
     ) -> CompositeType:
         log_prefix = "%s.%d.%d" % (self.full_name, self.version.major, self.version.minor)
         if self._cached_type is not None:
@@ -262,7 +264,7 @@ def read(
                 allow_unregulated_fixed_port_id=allow_unregulated_fixed_port_id,
             )
 
-            _parser.parse(self.text, builder)
+            _parser.parse(self.text, builder, strict=strict)
 
             self._cached_type = builder.finalize()
             _logger.info(

pydsdl/_namespace.py

@@ -74,6 +74,8 @@ def read_namespace(
     print_output_handler: PrintOutputHandler | None = None,
     allow_unregulated_fixed_port_id: bool = False,
     allow_root_namespace_name_collision: bool = True,
+    *,
+    strict: bool = False,
 ) -> list[_serializable.CompositeType]:
     """
     This function is a main entry point for the library.
@@ -103,6 +105,8 @@ def read_namespace(
              the same root namespace name multiple times in the lookup dirs. This will enable defining a namespace
              partially and let other entities define new messages or new sub-namespaces in the same root namespace.
 
+    :param strict: Reject features that are not [yet] part of the Cyphal Specification.
+
     :return: A list of :class:`pydsdl.CompositeType` found under the `root_namespace_directory` and sorted
              lexicographically by full data type name, then by major version (newest version first), then by minor
              version (newest version first). The ordering guarantee allows the caller to always find the newest version
@@ -131,7 +135,11 @@ def read_namespace(
         _logger.debug(_LOG_LIST_ITEM_PREFIX + str(x))
 
     return _complete_read_function(
-        target_dsdl_definitions, lookup_directories_path_list, print_output_handler, allow_unregulated_fixed_port_id
+        target_dsdl_definitions,
+        lookup_directories_path_list,
+        print_output_handler,
+        allow_unregulated_fixed_port_id=allow_unregulated_fixed_port_id,
+        strict=strict,
     ).direct
 
 
@@ -142,6 +150,8 @@ def read_files(
     lookup_directories: None | Path | str | Iterable[Path | str] = None,
     print_output_handler: PrintOutputHandler | None = None,
     allow_unregulated_fixed_port_id: bool = False,
+    *,
+    strict: bool = False,
 ) -> tuple[list[_serializable.CompositeType], list[_serializable.CompositeType]]:
     """
     This function is a main entry point for the library.
@@ -232,6 +242,8 @@ def read_files(
         This is a dangerous feature that must not be used unless you understand the risks.
         Please read https://opencyphal.org/guide.
 
+    :param strict: Reject features that are not [yet] part of the Cyphal Specification.
+
     :return: A Tuple of lists of :class:`pydsdl.CompositeType`. The first index in the Tuple are the types parsed from
         the ``dsdl_files`` argument. The second index are types that the target ``dsdl_files`` utilizes.
         A note for using these values to describe build dependencies: each :class:`pydsdl.CompositeType` has two
@@ -266,9 +278,12 @@ def read_files(
     )
 
     definitions = _complete_read_function(
-        target_dsdl_definitions, lookup_directories_path_list, print_output_handler, allow_unregulated_fixed_port_id
+        target_dsdl_definitions,
+        lookup_directories_path_list,
+        print_output_handler,
+        allow_unregulated_fixed_port_id=allow_unregulated_fixed_port_id,
+        strict=strict,
     )
-
     return (definitions.direct, definitions.transitive)
 
 
@@ -286,7 +301,9 @@ def _complete_read_function(
     target_dsdl_definitions: SortedFileList[ReadableDSDLFile],
     lookup_directories_path_list: list[Path],
     print_output_handler: PrintOutputHandler | None,
+    *,
     allow_unregulated_fixed_port_id: bool,
+    strict: bool,
 ) -> DSDLDefinitions:
 
     lookup_dsdl_definitions = _construct_dsdl_definitions_from_namespaces(lookup_directories_path_list)
@@ -307,7 +324,11 @@ def _complete_read_function(
     # This is the biggie. All the rest of the wrangling is just to get to this point. This will take the
     # most time and memory.
     definitions = read_definitions(
-        target_dsdl_definitions, lookup_dsdl_definitions, print_output_handler, allow_unregulated_fixed_port_id
+        target_dsdl_definitions,
+        lookup_dsdl_definitions,
+        print_output_handler,
+        allow_unregulated_fixed_port_id=allow_unregulated_fixed_port_id,
+        strict=strict,
     )
 
     # Note that we check for collisions in the read namespace only.

pydsdl/_namespace_reader.py

@@ -19,12 +19,14 @@
 def _read_definitions(
     target_definitions: SortedFileList[ReadableDSDLFile],
     lookup_definitions: SortedFileList[ReadableDSDLFile],
+    *,
     print_output_handler: PrintOutputHandler | None,
     allow_unregulated_fixed_port_id: bool,
+    strict: bool,
     direct: set[CompositeType],
     transitive: set[CompositeType],
     file_pool: dict[Path, ReadableDSDLFile],
-    level: int,
+    level: int = 0,
 ) -> None:
     """
     Don't look at me! I'm hideous!
@@ -87,12 +89,13 @@ def print_handler(file: Path, line: int, message: str) -> None:
             _read_definitions(
                 dsdl_file_sort(_pending_definitions),
                 lookup_definitions,
-                print_output_handler,
-                allow_unregulated_fixed_port_id,
-                direct,
-                transitive,
-                file_pool,
-                level + 1,
+                print_output_handler=print_output_handler,
+                allow_unregulated_fixed_port_id=allow_unregulated_fixed_port_id,
+                strict=strict,
+                direct=direct,
+                transitive=transitive,
+                file_pool=file_pool,
+                level=level + 1,
             )
             _pending_definitions.clear()
 
@@ -116,6 +119,8 @@ def read_definitions(
     lookup_definitions: SortedFileList[ReadableDSDLFile],
     print_output_handler: PrintOutputHandler | None,
     allow_unregulated_fixed_port_id: bool,
+    *,
+    strict: bool = False,
 ) -> DSDLDefinitions:
     """
     Given a set of DSDL files, this method reads the text and invokes the parser for each and for any files found in the
@@ -125,6 +130,7 @@ def read_definitions(
     :param lookup_definitions:              List of definitions available for referring to.
     :param print_output_handler:            Used for @print and for diagnostics: (line_number, text) -> None.
     :param allow_unregulated_fixed_port_id: Do not complain about fixed unregulated port IDs.
+    :param strict:                          Reject features that are not part of the Specification.
     :return: The data type representation.
     :raises InvalidDefinitionError: If a dependency is missing.
     :raises InternalError: If an unexpected error occurs.
@@ -135,12 +141,12 @@ def read_definitions(
     _read_definitions(
         target_definitions,
         lookup_definitions,
-        print_output_handler,
-        allow_unregulated_fixed_port_id,
-        _direct,
-        _transitive,
-        _file_pool,
-        0,
+        print_output_handler=print_output_handler,
+        allow_unregulated_fixed_port_id=allow_unregulated_fixed_port_id,
+        strict=strict,
+        direct=_direct,
+        transitive=_transitive,
+        file_pool=_file_pool,
     )
     return DSDLDefinitions(
         dsdl_file_sort(_direct),
@@ -186,6 +192,7 @@ def _unittest_namespace_reader_read_definitions_multiple(temp_dsdl_factory) -> N
     assert len(definitions.transitive) == 2
 
 
+# noinspection PyProtectedMember
 def _unittest_namespace_reader_read_definitions_multiple_no_load(temp_dsdl_factory) -> None:  # type: ignore
     """
     Ensure that the loader does not load files that are not in the transitive closure of the target files.

pydsdl/_parser.py

@@ -21,12 +21,12 @@ class DSDLSyntaxError(_error.InvalidDefinitionError):
     pass
 
 
-def parse(text: str, statement_stream_processor: "StatementStreamProcessor") -> None:
+def parse(text: str, statement_stream_processor: "StatementStreamProcessor", *, strict: bool) -> None:
     """
     The entry point of the parser. As the text is being parsed, the parser invokes appropriate
     methods in the statement stream processor.
     """
-    pr = _ParseTreeProcessor(statement_stream_processor)
+    pr = _ParseTreeProcessor(statement_stream_processor, strict=strict)
     try:
         pr.visit(_get_grammar().parse(text))  # type: ignore
     except _error.FrontendError as ex:
@@ -134,12 +134,13 @@ class _ParseTreeProcessor(parsimonious.NodeVisitor):
     # Beware that those might be propagated from recursive parser instances!
     unwrapped_exceptions = (_error.FrontendError, SystemError, MemoryError, SystemExit)  # type: ignore
 
-    def __init__(self, statement_stream_processor: StatementStreamProcessor):
+    def __init__(self, statement_stream_processor: StatementStreamProcessor, *, strict: bool):
         assert isinstance(statement_stream_processor, StatementStreamProcessor)
         self._statement_stream_processor = statement_stream_processor  # type: StatementStreamProcessor
         self._current_line_number = 1  # Lines are numbered from one
         self._comment = ""
         self._comment_is_header = True
+        self._strict = bool(strict)
         super().__init__()
 
     @property
@@ -269,9 +270,13 @@ def visit_type_primitive_boolean(self, _n: _Node, _c: _Children) -> _serializabl
         return _serializable.BooleanType()
 
     def visit_type_primitive_byte(self, _n: _Node, _c: _Children) -> _serializable.PrimitiveType:
+        if self._strict:
+            raise _error.InvalidDefinitionError("byte is a non-standard extension unavailable in strict mode")
         return _serializable.ByteType()
 
     def visit_type_primitive_utf8(self, _n: _Node, _c: _Children) -> _serializable.PrimitiveType:
+        if self._strict:
+            raise _error.InvalidDefinitionError("utf8 is a non-standard extension unavailable in strict mode")
         return _serializable.UTF8Type()
 
     def visit_type_primitive_truncated(self, _n: _Node, children: _Children) -> _serializable.PrimitiveType:

pydsdl/_serializable/_composite.py

@@ -14,7 +14,6 @@
 from ._attribute import Attribute, Field, PaddingField, Constant
 from ._name import check_name, InvalidNameError
 from ._primitive import PrimitiveType, UnsignedIntegerType
-from ._void import VoidType
 
 Version = typing.NamedTuple("Version", [("major", int), ("minor", int)])
 

pydsdl/_test.py

@@ -7,11 +7,11 @@
 
 from __future__ import annotations
 import tempfile
-from typing import Sequence, Type, Iterable
+from typing import Sequence, Type, Iterable, Any
 from pathlib import Path
 from textwrap import dedent
 import pytest  # This is only safe to import in test files!
-from . import _data_type_builder
+from . import _data_type_builder, InvalidDefinitionError
 from . import _expression
 from . import _error
 from . import _parser
@@ -67,13 +67,16 @@ def is_fs_case_sensitive(self) -> bool:
 
 
 def parse_definition(
-    definition: _dsdl_definition.DSDLDefinition, lookup_definitions: Sequence[_dsdl_definition.DSDLDefinition]
+    definition: _dsdl_definition.DSDLDefinition,
+    lookup_definitions: Sequence[_dsdl_definition.DSDLDefinition],
+    **kwargs: Any,
 ) -> _serializable.CompositeType:
     return definition.read(
         lookup_definitions,
         [],
         print_output_handler=lambda line, text: print("Output from line %d:" % line, text),
         allow_unregulated_fixed_port_id=False,
+        **kwargs,
     )
 
 
@@ -1986,6 +1989,22 @@ def _unittest_dsdl_parser_utf8_bytes(wrkspc: Workspace) -> None:
     assert t.string_like
 
 
+def _unittest_dsdl_parser_strict_mode(wrkspc: Workspace) -> None:
+    from pytest import raises
+
+    name = "ns/A.1.0.dsdl"
+    src = dedent(
+        r"""
+        byte[<=10] a
+        utf8[<=10] b
+        @sealed
+        """
+    )
+    _ = parse_definition(wrkspc.parse_new(name, src), [])  # success by default
+    with raises(InvalidDefinitionError, match="(?i).*strict.*"):
+        parse_definition(wrkspc.parse_new(name, src), [], strict=True)
+
+
 def _unittest_dsdl_parser_expressions(wrkspc: Workspace) -> None:
     from pytest import raises
 

setup.cfg

@@ -119,7 +119,9 @@ disable=
     too-many-public-methods,
     consider-using-f-string,
     unspecified-encoding,
-    use-implicit-booleaness-not-comparison
+    use-implicit-booleaness-not-comparison,
+    too-many-arguments,
+    too-many-locals
 
 [pylint.REPORTS]
 output-format=colorized