mirror of
https://github.com/LadybirdBrowser/ladybird
synced 2026-05-11 01:22:43 +02:00
6fc06afc81
Add a simple python3 WebIDL parser for the subset of IDL we need for code generation. For validation, the parser structure is intentionally very close to the existing C++ parser, so the control flow and parsing mechanics stay easy to compare while we continue porting IDL-related generators. This parser currently focuses on the constructs needed by GenerateWindowOrWorkerInterfaces, but is meant to serve as a base for future IDL generator ports as well.
563 lines
18 KiB
Python
563 lines
18 KiB
Python
# Copyright (c) 2026-present, the Ladybird developers.
|
|
#
|
|
# SPDX-License-Identifier: BSD-2-Clause
|
|
|
|
from dataclasses import dataclass
|
|
from dataclasses import field
|
|
from pathlib import Path
|
|
from typing import Dict
|
|
from typing import List
|
|
from typing import Optional
|
|
|
|
from Utils.lexer import Lexer
|
|
|
|
|
|
@dataclass
|
|
class Constant:
|
|
declaration: str
|
|
|
|
|
|
@dataclass
|
|
class SpecialOperation:
|
|
identifier_type: str
|
|
declaration: str
|
|
|
|
|
|
@dataclass
|
|
class Interface:
|
|
name: str
|
|
path: Path
|
|
extended_attributes: Dict[str, str] = field(default_factory=dict)
|
|
is_namespace: bool = False
|
|
is_callback_interface: bool = False
|
|
parent_name: str = ""
|
|
constants: List[Constant] = field(default_factory=list)
|
|
named_property_getter: Optional[SpecialOperation] = None
|
|
indexed_property_getter: Optional[SpecialOperation] = None
|
|
implemented_name: str = ""
|
|
namespaced_name: str = ""
|
|
constructor_class: str = ""
|
|
prototype_class: str = ""
|
|
namespace_class: str = ""
|
|
|
|
def supports_named_properties(self) -> bool:
|
|
return self.named_property_getter is not None
|
|
|
|
def finalize(self) -> None:
|
|
legacy_namespace = self.extended_attributes.get("LegacyNamespace")
|
|
if legacy_namespace:
|
|
self.namespaced_name = f"{legacy_namespace}.{self.name}"
|
|
else:
|
|
self.namespaced_name = self.name
|
|
|
|
self.implemented_name = self.extended_attributes.get("ImplementedAs", self.name)
|
|
self.constructor_class = f"{self.implemented_name}Constructor"
|
|
self.prototype_class = f"{self.implemented_name}Prototype"
|
|
self.namespace_class = f"{self.name}Namespace"
|
|
|
|
|
|
@dataclass
|
|
class Module:
|
|
path: Path
|
|
interface: Optional[Interface] = None
|
|
|
|
|
|
@dataclass
|
|
class NestingState:
|
|
parentheses_depth: int = 0
|
|
bracket_depth: int = 0
|
|
brace_depth: int = 0
|
|
angle_depth: int = 0
|
|
|
|
def is_at_top_level(self) -> bool:
|
|
return not any((self.parentheses_depth, self.bracket_depth, self.brace_depth, self.angle_depth))
|
|
|
|
def update_for_character(self, character: str) -> None:
|
|
if character == "(":
|
|
self.parentheses_depth += 1
|
|
elif character == ")":
|
|
self.parentheses_depth = max(self.parentheses_depth - 1, 0)
|
|
elif character == "[":
|
|
self.bracket_depth += 1
|
|
elif character == "]":
|
|
self.bracket_depth = max(self.bracket_depth - 1, 0)
|
|
elif character == "{":
|
|
self.brace_depth += 1
|
|
elif character == "}":
|
|
self.brace_depth = max(self.brace_depth - 1, 0)
|
|
elif character == "<":
|
|
self.angle_depth += 1
|
|
elif character == ">":
|
|
self.angle_depth = max(self.angle_depth - 1, 0)
|
|
|
|
|
|
class ParseError(RuntimeError):
|
|
pass
|
|
|
|
|
|
def parse_module(path: Path, contents: str) -> Module:
|
|
return Parser(path, contents).parse()
|
|
|
|
|
|
class Parser:
|
|
def __init__(self, path: Path, contents: str) -> None:
|
|
self.path = path
|
|
self.contents = contents
|
|
self.lexer = Lexer(contents)
|
|
|
|
def parse(self) -> Module:
|
|
module = Module(path=self.path)
|
|
|
|
self.consume_whitespace()
|
|
while not self.lexer.is_eof():
|
|
extended_attributes: Dict[str, str] = {}
|
|
if self.lexer.consume_specific("["):
|
|
extended_attributes = self.parse_extended_attributes()
|
|
|
|
if self.next_is_keyword("dictionary") or self.next_is_keyword("partial dictionary"):
|
|
self.skip_braced_declaration()
|
|
elif self.next_is_keyword("enum"):
|
|
self.skip_braced_declaration()
|
|
elif self.next_is_keyword("typedef"):
|
|
self.consume_keyword("typedef")
|
|
self.consume_statement_text()
|
|
elif self.next_is_keyword("partial interface mixin"):
|
|
self.skip_braced_declaration()
|
|
elif self.next_is_keyword("partial interface"):
|
|
self.skip_braced_declaration()
|
|
elif self.next_is_keyword("interface mixin"):
|
|
self.skip_braced_declaration()
|
|
elif self.next_is_keyword("partial namespace"):
|
|
self.skip_braced_declaration()
|
|
elif self.next_is_keyword("callback interface"):
|
|
module.interface = self.set_or_check_module_interface(
|
|
module.interface,
|
|
self.parse_interface(extended_attributes, is_callback_interface=True),
|
|
)
|
|
elif self.next_is_keyword("callback"):
|
|
self.consume_keyword("callback")
|
|
self.consume_statement_text()
|
|
elif self.next_is_keyword("namespace"):
|
|
module.interface = self.set_or_check_module_interface(
|
|
module.interface,
|
|
self.parse_interface(extended_attributes, is_namespace=True),
|
|
)
|
|
elif self.next_is_keyword("interface"):
|
|
module.interface = self.set_or_check_module_interface(
|
|
module.interface,
|
|
self.parse_interface(extended_attributes),
|
|
)
|
|
else:
|
|
self.parse_includes_statement()
|
|
|
|
self.consume_whitespace()
|
|
|
|
if module.interface is None:
|
|
self.raise_parse_error("did not find an interface, callback interface, or namespace")
|
|
|
|
return module
|
|
|
|
def set_or_check_module_interface(
|
|
self, existing_interface: Optional[Interface], candidate_interface: Interface
|
|
) -> Interface:
|
|
if existing_interface is None:
|
|
return candidate_interface
|
|
|
|
if not existing_interface.is_namespace and existing_interface.name == candidate_interface.name:
|
|
return existing_interface
|
|
|
|
self.raise_parse_error(
|
|
"encountered multiple interface, callback interface, or namespace declarations in one file"
|
|
)
|
|
|
|
def parse_includes_statement(self) -> None:
|
|
self.parse_identifier_ending_with_space()
|
|
self.consume_whitespace()
|
|
|
|
if not self.next_is_keyword("includes"):
|
|
self.raise_parse_error("expected a declaration or an includes statement")
|
|
|
|
self.consume_keyword("includes")
|
|
self.consume_whitespace()
|
|
self.parse_identifier_ending_with_space_or(";")
|
|
self.consume_whitespace()
|
|
self.assert_specific(";")
|
|
|
|
def parse_interface(
|
|
self,
|
|
extended_attributes: Dict[str, str],
|
|
is_namespace: bool = False,
|
|
is_callback_interface: bool = False,
|
|
) -> Interface:
|
|
if is_callback_interface:
|
|
self.consume_keyword("callback")
|
|
self.consume_whitespace()
|
|
self.consume_keyword("interface")
|
|
elif is_namespace:
|
|
self.consume_keyword("namespace")
|
|
else:
|
|
self.consume_keyword("interface")
|
|
|
|
self.consume_whitespace()
|
|
|
|
interface = Interface(
|
|
name=self.parse_identifier_ending_with_space_or(":", "{"),
|
|
path=self.path,
|
|
extended_attributes=extended_attributes,
|
|
is_namespace=is_namespace,
|
|
is_callback_interface=is_callback_interface,
|
|
)
|
|
|
|
self.consume_whitespace()
|
|
if not is_namespace and self.lexer.consume_specific(":"):
|
|
self.consume_whitespace()
|
|
interface.parent_name = self.parse_identifier_ending_with_space_or("{")
|
|
self.consume_whitespace()
|
|
|
|
body_text = self.consume_braced_block()
|
|
self.consume_whitespace()
|
|
self.assert_specific(";")
|
|
|
|
if not is_namespace:
|
|
self.parse_interface_body(interface, body_text)
|
|
|
|
interface.finalize()
|
|
return interface
|
|
|
|
def parse_interface_body(self, interface: Interface, body_text: str) -> None:
|
|
for statement in split_top_level_statements(remove_line_comments(body_text)):
|
|
if not statement:
|
|
continue
|
|
|
|
stripped_statement = strip_leading_extended_attributes(statement).strip()
|
|
if not stripped_statement:
|
|
continue
|
|
|
|
if stripped_statement.startswith("const "):
|
|
interface.constants.append(Constant(stripped_statement))
|
|
continue
|
|
|
|
if stripped_statement.startswith("getter "):
|
|
identifier_type = parse_special_operation_identifier_type(stripped_statement)
|
|
special_operation = SpecialOperation(identifier_type=identifier_type, declaration=stripped_statement)
|
|
|
|
if identifier_type == "DOMString":
|
|
interface.named_property_getter = special_operation
|
|
elif identifier_type == "unsigned long":
|
|
interface.indexed_property_getter = special_operation
|
|
else:
|
|
self.raise_parse_error(
|
|
f"named/indexed property getter must use DOMString or unsigned long, got '{identifier_type}'"
|
|
)
|
|
|
|
def parse_extended_attributes(self) -> Dict[str, str]:
|
|
extended_attributes: Dict[str, str] = {}
|
|
|
|
while True:
|
|
self.consume_whitespace()
|
|
if self.lexer.consume_specific("]"):
|
|
break
|
|
|
|
name = self.parse_identifier_ending_with_space_or("]", "=", ",")
|
|
value = ""
|
|
if self.lexer.consume_specific("="):
|
|
value = self.consume_extended_attribute_value().strip()
|
|
|
|
extended_attributes[name] = value
|
|
self.lexer.consume_specific(",")
|
|
|
|
self.consume_whitespace()
|
|
return extended_attributes
|
|
|
|
def consume_extended_attribute_value(self) -> str:
|
|
start = self.lexer.tell()
|
|
parentheses_depth = 0
|
|
|
|
while not self.lexer.is_eof():
|
|
character = self.lexer.peek()
|
|
if character == "(":
|
|
parentheses_depth += 1
|
|
elif character == ")":
|
|
if parentheses_depth > 0:
|
|
parentheses_depth -= 1
|
|
elif parentheses_depth == 0 and character in (",", "]"):
|
|
break
|
|
|
|
self.lexer.consume()
|
|
|
|
return self.contents[start : self.lexer.tell()]
|
|
|
|
def skip_braced_declaration(self) -> None:
|
|
while not self.lexer.is_eof() and self.lexer.peek() != "{":
|
|
self.lexer.consume()
|
|
|
|
if self.lexer.is_eof():
|
|
self.raise_parse_error("expected '{' while skipping declaration")
|
|
|
|
self.consume_braced_block()
|
|
self.consume_whitespace()
|
|
self.assert_specific(";")
|
|
|
|
def consume_braced_block(self) -> str:
|
|
self.assert_specific("{")
|
|
|
|
start = self.lexer.tell()
|
|
brace_depth = 1
|
|
active_quote = ""
|
|
|
|
while not self.lexer.is_eof():
|
|
character = self.lexer.consume()
|
|
|
|
if active_quote:
|
|
if character == active_quote:
|
|
active_quote = ""
|
|
continue
|
|
|
|
if character == "/" and self.lexer.peek() == "/":
|
|
self.lexer.ignore_until("\n")
|
|
if self.lexer.peek() == "\n":
|
|
self.lexer.ignore()
|
|
continue
|
|
|
|
if character in ('"', "'"):
|
|
active_quote = character
|
|
continue
|
|
|
|
if character == "{":
|
|
brace_depth += 1
|
|
elif character == "}":
|
|
brace_depth -= 1
|
|
if brace_depth == 0:
|
|
end = self.lexer.tell() - 1
|
|
return self.contents[start:end]
|
|
|
|
self.raise_parse_error("unterminated declaration body")
|
|
|
|
def consume_statement_text(self) -> str:
|
|
start = self.lexer.tell()
|
|
nesting_state = NestingState()
|
|
active_quote = ""
|
|
|
|
while not self.lexer.is_eof():
|
|
character = self.lexer.consume()
|
|
|
|
if active_quote:
|
|
if character == active_quote:
|
|
active_quote = ""
|
|
continue
|
|
|
|
if character == "/" and self.lexer.peek() == "/":
|
|
self.lexer.ignore_until("\n")
|
|
if self.lexer.peek() == "\n":
|
|
self.lexer.ignore()
|
|
continue
|
|
|
|
if character in ('"', "'"):
|
|
active_quote = character
|
|
continue
|
|
|
|
if character == ";" and nesting_state.is_at_top_level():
|
|
end = self.lexer.tell() - 1
|
|
return self.contents[start:end]
|
|
nesting_state.update_for_character(character)
|
|
|
|
self.raise_parse_error("unterminated statement")
|
|
|
|
def parse_identifier_ending_with_space(self) -> str:
|
|
return self.parse_identifier_ending_with_space_or()
|
|
|
|
def parse_identifier_ending_with_space_or(self, *terminators: str) -> str:
|
|
identifier = self.lexer.consume_until(lambda character: character.isspace() or character in terminators)
|
|
return identifier.lstrip("_")
|
|
|
|
def consume_whitespace(self) -> None:
|
|
consumed = True
|
|
while consumed:
|
|
consumed = False
|
|
|
|
whitespace = self.lexer.consume_while(str.isspace)
|
|
if whitespace:
|
|
consumed = True
|
|
|
|
if self.lexer.consume_specific("//"):
|
|
self.lexer.ignore_until("\n")
|
|
if self.lexer.peek() == "\n":
|
|
self.lexer.ignore()
|
|
consumed = True
|
|
|
|
def next_is_keyword(self, keyword: str) -> bool:
|
|
if not self.lexer.next_is(keyword):
|
|
return False
|
|
|
|
end = self.lexer.tell() + len(keyword)
|
|
if end >= len(self.contents):
|
|
return True
|
|
|
|
trailing_character = self.contents[end]
|
|
return trailing_character.isspace() or trailing_character in "{([;:"
|
|
|
|
def consume_keyword(self, keyword: str) -> None:
|
|
if not self.next_is_keyword(keyword):
|
|
self.raise_parse_error(f"expected '{keyword}'")
|
|
self.lexer.ignore(len(keyword))
|
|
|
|
def assert_specific(self, expected_character: str) -> None:
|
|
if not self.lexer.consume_specific(expected_character):
|
|
self.raise_parse_error(f"expected '{expected_character}'")
|
|
|
|
def raise_parse_error(self, message: str) -> None:
|
|
line_number = 1
|
|
column_number = 1
|
|
|
|
for index, character in enumerate(self.contents):
|
|
if index == self.lexer.tell():
|
|
break
|
|
if character == "\n":
|
|
line_number += 1
|
|
column_number = 1
|
|
else:
|
|
column_number += 1
|
|
|
|
raise ParseError(f"{self.path}:{line_number}:{column_number}: error: {message}")
|
|
|
|
|
|
def remove_line_comments(text: str) -> str:
|
|
lines = []
|
|
for line in text.splitlines():
|
|
comment_index = line.find("//")
|
|
if comment_index != -1:
|
|
line = line[:comment_index]
|
|
lines.append(line)
|
|
return "\n".join(lines)
|
|
|
|
|
|
def split_top_level_statements(text: str) -> List[str]:
|
|
statements: List[str] = []
|
|
start = 0
|
|
nesting_state = NestingState()
|
|
active_quote = ""
|
|
|
|
for index, character in enumerate(text):
|
|
if active_quote:
|
|
if character == active_quote:
|
|
active_quote = ""
|
|
continue
|
|
|
|
if character in ('"', "'"):
|
|
active_quote = character
|
|
continue
|
|
|
|
if character == ";" and nesting_state.is_at_top_level():
|
|
statements.append(text[start:index].strip())
|
|
start = index + 1
|
|
continue
|
|
|
|
nesting_state.update_for_character(character)
|
|
|
|
trailing_text = text[start:].strip()
|
|
if trailing_text:
|
|
statements.append(trailing_text)
|
|
|
|
return statements
|
|
|
|
|
|
def strip_leading_extended_attributes(text: str) -> str:
|
|
remaining_text = text.lstrip()
|
|
while remaining_text.startswith("["):
|
|
closing_bracket = find_matching_closer(remaining_text, 0, "[", "]")
|
|
remaining_text = remaining_text[closing_bracket + 1 :].lstrip()
|
|
return remaining_text
|
|
|
|
|
|
def parse_special_operation_identifier_type(statement: str) -> str:
|
|
parenthesized_text = extract_final_parenthesized_group(statement)
|
|
parameter_text = split_top_level_items(parenthesized_text, ",")[0].strip()
|
|
parameter_text = strip_leading_extended_attributes(parameter_text)
|
|
|
|
last_space = find_last_top_level_whitespace(parameter_text)
|
|
if last_space == -1:
|
|
raise ParseError(f"could not determine getter identifier type from '{statement}'")
|
|
|
|
return normalize_whitespace(parameter_text[:last_space])
|
|
|
|
|
|
def extract_final_parenthesized_group(text: str) -> str:
|
|
stripped_text = text.rstrip()
|
|
if not stripped_text.endswith(")"):
|
|
raise ParseError(f"expected a trailing parameter list in '{text}'")
|
|
|
|
closing_index = len(stripped_text) - 1
|
|
opening_index = find_matching_opener(stripped_text, closing_index, "(", ")")
|
|
return stripped_text[opening_index + 1 : closing_index]
|
|
|
|
|
|
def split_top_level_items(text: str, delimiter: str) -> List[str]:
|
|
items: List[str] = []
|
|
start = 0
|
|
nesting_state = NestingState()
|
|
active_quote = ""
|
|
|
|
for index, character in enumerate(text):
|
|
if active_quote:
|
|
if character == active_quote:
|
|
active_quote = ""
|
|
continue
|
|
|
|
if character in ('"', "'"):
|
|
active_quote = character
|
|
continue
|
|
|
|
if character == delimiter and nesting_state.is_at_top_level():
|
|
items.append(text[start:index])
|
|
start = index + 1
|
|
continue
|
|
|
|
nesting_state.update_for_character(character)
|
|
|
|
items.append(text[start:])
|
|
return items
|
|
|
|
|
|
def find_last_top_level_whitespace(text: str) -> int:
|
|
last_space = -1
|
|
nesting_state = NestingState()
|
|
|
|
for index, character in enumerate(text):
|
|
if character.isspace() and nesting_state.is_at_top_level():
|
|
last_space = index
|
|
nesting_state.update_for_character(character)
|
|
|
|
return last_space
|
|
|
|
|
|
def find_matching_closer(text: str, start_index: int, opener: str, closer: str) -> int:
|
|
depth = 0
|
|
for index in range(start_index, len(text)):
|
|
character = text[index]
|
|
if character == opener:
|
|
depth += 1
|
|
elif character == closer:
|
|
depth -= 1
|
|
if depth == 0:
|
|
return index
|
|
|
|
raise ParseError(f"unterminated '{opener}' in '{text}'")
|
|
|
|
|
|
def find_matching_opener(text: str, start_index: int, opener: str, closer: str) -> int:
|
|
depth = 0
|
|
for index in range(start_index, -1, -1):
|
|
character = text[index]
|
|
if character == closer:
|
|
depth += 1
|
|
elif character == opener:
|
|
depth -= 1
|
|
if depth == 0:
|
|
return index
|
|
|
|
raise ParseError(f"unterminated '{opener}' in '{text}'")
|
|
|
|
|
|
def normalize_whitespace(text: str) -> str:
|
|
return " ".join(text.split())
|