X-Git-Url: https://git.madduck.net/etc/vim.git/blobdiff_plain/b85c345dbc70302610c05cc97fe30f5790fadc79..2848e2e1d6527d6031ea020cd991fd73e52c4a0b:/black.py diff --git a/black.py b/black.py index 728b084..910a0ed 100644 --- a/black.py +++ b/black.py @@ -1,6 +1,7 @@ +import ast import asyncio -from asyncio.base_events import BaseEventLoop from concurrent.futures import Executor, ProcessPoolExecutor +from contextlib import contextmanager from datetime import datetime from enum import Enum from functools import lru_cache, partial, wraps @@ -16,6 +17,7 @@ import signal import sys import tempfile import tokenize +import traceback from typing import ( Any, Callable, @@ -140,6 +142,8 @@ class Feature(Enum): # set for every version of python. ASYNC_IDENTIFIERS = 6 ASYNC_KEYWORDS = 7 + ASSIGNMENT_EXPRESSIONS = 8 + POS_ONLY_ARGUMENTS = 9 VERSION_TO_FEATURES: Dict[TargetVersion, Set[Feature]] = { @@ -174,6 +178,8 @@ VERSION_TO_FEATURES: Dict[TargetVersion, Set[Feature]] = { Feature.TRAILING_COMMA_IN_CALL, Feature.TRAILING_COMMA_IN_DEF, Feature.ASYNC_KEYWORDS, + Feature.ASSIGNMENT_EXPRESSIONS, + Feature.POS_ONLY_ARGUMENTS, }, } @@ -336,7 +342,7 @@ def read_pyproject_toml( "--quiet", is_flag=True, help=( - "Don't emit non-error messages to stderr. Errors are still emitted, " + "Don't emit non-error messages to stderr. Errors are still emitted; " "silence those with 2>/dev/null." ), ) @@ -467,8 +473,7 @@ def reformat_one( ) -> None: """Reformat a single file under `src` without spawning child processes. - If `quiet` is True, non-error messages are not output. `line_length`, - `write_back`, `fast` and `pyi` options are passed to + `fast`, `write_back`, and `mode` options are passed to :func:`format_file_in_place` or :func:`format_stdin_to_stdout`. """ try: @@ -524,6 +529,7 @@ def reformat_many( ) finally: shutdown(loop) + executor.shutdown() async def schedule_formatting( @@ -532,14 +538,14 @@ async def schedule_formatting( write_back: WriteBack, mode: FileMode, report: "Report", - loop: BaseEventLoop, + loop: asyncio.AbstractEventLoop, executor: Executor, ) -> None: """Run formatting of `sources` in parallel using the provided `executor`. (Use ProcessPoolExecutors for actual parallelism.) - `line_length`, `write_back`, `fast`, and `pyi` options are passed to + `write_back`, `fast`, and `mode` options are passed to :func:`format_file_in_place`. """ cache: Cache = {} @@ -608,7 +614,7 @@ def format_file_in_place( If `write_back` is DIFF, write a diff to stdout. If it is YES, write reformatted code to the file. - `line_length` and `fast` options are passed to :func:`format_file_contents`. + `mode` and `fast` options are passed to :func:`format_file_contents`. """ if src.suffix == ".pyi": mode = evolve(mode, is_pyi=True) @@ -629,9 +635,8 @@ def format_file_in_place( src_name = f"{src}\t{then} +0000" dst_name = f"{src}\t{now} +0000" diff_contents = diff(src_contents, dst_contents, src_name, dst_name) - if lock: - lock.acquire() - try: + + with lock or nullcontext(): f = io.TextIOWrapper( sys.stdout.buffer, encoding=encoding, @@ -640,9 +645,7 @@ def format_file_in_place( ) f.write(diff_contents) f.detach() - finally: - if lock: - lock.release() + return True @@ -686,7 +689,7 @@ def format_file_contents( If `fast` is False, additionally confirm that the reformatted code is valid by calling :func:`assert_equivalent` and :func:`assert_stable` on it. - `line_length` is passed to :func:`format_str`. + `mode` is passed to :func:`format_str`. """ if src_contents.strip() == "": raise NothingChanged @@ -704,10 +707,11 @@ def format_file_contents( def format_str(src_contents: str, *, mode: FileMode) -> FileContent: """Reformat a string and return new contents. - `line_length` determines how many characters per line are allowed. + `mode` determines formatting options, such as how many characters per line are + allowed. """ src_node = lib2to3_parse(src_contents.lstrip(), mode.target_versions) - dst_contents = "" + dst_contents = [] future_imports = get_future_imports(src_node) if mode.target_versions: versions = mode.target_versions @@ -730,15 +734,15 @@ def format_str(src_contents: str, *, mode: FileMode) -> FileContent: } for current_line in lines.visit(src_node): for _ in range(after): - dst_contents += str(empty_line) + dst_contents.append(str(empty_line)) before, after = elt.maybe_empty_lines(current_line) for _ in range(before): - dst_contents += str(empty_line) + dst_contents.append(str(empty_line)) for line in split_line( current_line, line_length=mode.line_length, features=split_line_features ): - dst_contents += str(line) - return dst_contents + dst_contents.append(str(line)) + return "".join(dst_contents) def decode_bytes(src: bytes) -> Tuple[FileContent, Encoding, NewLine]: @@ -933,6 +937,7 @@ MATH_OPERATORS = { token.DOUBLESTAR, } STARS = {token.STAR, token.DOUBLESTAR} +VARARGS_SPECIALS = STARS | {token.SLASH} VARARGS_PARENTS = { syms.arglist, syms.argument, # double star in arglist @@ -1060,7 +1065,7 @@ class BracketTracker: """Return True if there is an yet unmatched open bracket on the line.""" return bool(self.bracket_match) - def max_delimiter_priority(self, exclude: Iterable[LeafID] = ()) -> int: + def max_delimiter_priority(self, exclude: Iterable[LeafID] = ()) -> Priority: """Return the highest priority of a delimiter found on the line. Values are consistent with what `is_split_*_delimiter()` return. @@ -1068,7 +1073,7 @@ class BracketTracker: """ return max(v for k, v in self.delimiters.items() if k not in exclude) - def delimiter_count_with_priority(self, priority: int = 0) -> int: + def delimiter_count_with_priority(self, priority: Priority = 0) -> int: """Return the number of delimiters with the given `priority`. If no `priority` is passed, defaults to max priority on the line. @@ -1281,10 +1286,13 @@ class Line: try: last_leaf = self.leaves[-1] ignored_ids.add(id(last_leaf)) - if last_leaf.type == token.COMMA: - # When trailing commas are inserted by Black for consistency, comments - # after the previous last element are not moved (they don't have to, - # rendering will still be correct). So we ignore trailing commas. + if last_leaf.type == token.COMMA or ( + last_leaf.type == token.RPAR and not last_leaf.value + ): + # When trailing commas or optional parens are inserted by Black for + # consistency, comments after the previous last element are not moved + # (they don't have to, rendering will still be correct). So we ignore + # trailing commas and invisible. last_leaf = self.leaves[-2] ignored_ids.add(id(last_leaf)) except IndexError: @@ -1351,7 +1359,10 @@ class Line: bracket_depth = leaf.bracket_depth if bracket_depth == depth and leaf.type == token.COMMA: commas += 1 - if leaf.parent and leaf.parent.type == syms.arglist: + if leaf.parent and leaf.parent.type in { + syms.arglist, + syms.typedargslist, + }: commas += 1 break @@ -1378,7 +1389,23 @@ class Line: comment.prefix = "" return False - self.comments.setdefault(id(self.leaves[-1]), []).append(comment) + last_leaf = self.leaves[-1] + if ( + last_leaf.type == token.RPAR + and not last_leaf.value + and last_leaf.parent + and len(list(last_leaf.parent.leaves())) <= 3 + and not is_type_comment(comment) + ): + # Comments on an optional parens wrapping a single leaf should belong to + # the wrapped node except if it's a type comment. Pinning the comment like + # this avoids unstable formatting caused by comment migration. + if len(self.leaves) < 2: + comment.type = STANDALONE_COMMENT + comment.prefix = "" + return False + last_leaf = self.leaves[-2] + self.comments.setdefault(id(last_leaf), []).append(comment) return True def comments_after(self, leaf: Leaf) -> List[Leaf]: @@ -1621,6 +1648,19 @@ class LineGenerator(Visitor[Line]): node.children[2].value = "" yield from super().visit_default(node) + def visit_factor(self, node: Node) -> Iterator[Line]: + """Force parentheses between a unary op and a binary power: + + -2 ** 8 -> -(2 ** 8) + """ + child = node.children[1] + if child.type == syms.power and len(child.children) == 3: + lpar = Leaf(token.LPAR, "(") + rpar = Leaf(token.RPAR, ")") + index = child.remove() or 0 + node.insert_child(index, Node(syms.atom, [lpar, child, rpar])) + yield from self.visit_default(node) + def visit_INDENT(self, node: Node) -> Iterator[Line]: """Increase indentation level, maybe yield a line.""" # In blib2to3 INDENT never holds comments. @@ -1810,7 +1850,7 @@ def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str: # noqa: C901 # that, too. return prevp.prefix - elif prevp.type in STARS: + elif prevp.type in VARARGS_SPECIALS: if is_vararg(prevp, within=VARARGS_PARENTS | UNPACKING_PARENTS): return NO @@ -1900,7 +1940,7 @@ def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str: # noqa: C901 if not prevp or prevp.type == token.LPAR: return NO - elif prev.type in {token.EQUAL} | STARS: + elif prev.type in {token.EQUAL} | VARARGS_SPECIALS: return NO elif p.type == syms.decorator: @@ -2034,7 +2074,7 @@ def container_of(leaf: Leaf) -> LN: return container -def is_split_after_delimiter(leaf: Leaf, previous: Optional[Leaf] = None) -> int: +def is_split_after_delimiter(leaf: Leaf, previous: Optional[Leaf] = None) -> Priority: """Return the priority of the `leaf` delimiter, given a line break after it. The delimiter priorities returned here are from those delimiters that would @@ -2048,7 +2088,7 @@ def is_split_after_delimiter(leaf: Leaf, previous: Optional[Leaf] = None) -> int return 0 -def is_split_before_delimiter(leaf: Leaf, previous: Optional[Leaf] = None) -> int: +def is_split_before_delimiter(leaf: Leaf, previous: Optional[Leaf] = None) -> Priority: """Return the priority of the `leaf` delimiter, given a line break before it. The delimiter priorities returned here are from those delimiters that would @@ -2487,9 +2527,13 @@ def bracket_split_build_line( if leaves: # Since body is a new indent level, remove spurious leading whitespace. normalize_prefix(leaves[0], inside_brackets=True) - # Ensure a trailing comma for imports, but be careful not to add one after - # any comments. - if original.is_import: + # Ensure a trailing comma for imports and standalone function arguments, but + # be careful not to add one after any comments. + no_commas = original.is_def and not any( + l.type == token.COMMA for l in leaves + ) + + if original.is_import or no_commas: for i in range(len(leaves) - 1, -1, -1): if leaves[i].type == STANDALONE_COMMENT: continue @@ -2728,7 +2772,15 @@ def normalize_string_quotes(leaf: Leaf) -> None: new_body = sub_twice(escaped_orig_quote, rf"\1\2{orig_quote}", new_body) new_body = sub_twice(unescaped_new_quote, rf"\1\\{new_quote}", new_body) if "f" in prefix.casefold(): - matches = re.findall(r"[^{]\{(.*?)\}[^}]", new_body) + matches = re.findall( + r""" + (?:[^{]|^)\{ # start of the string or a non-{ followed by a single { + ([^{].*?) # contents of the brackets except if begins with {{ + \}(?:[^}]|$) # A } followed by end of the string or a non-} + """, + new_body, + re.VERBOSE, + ) for m in matches: if "\\" in str(m): # Do not introduce backslashes in interpolated expressions @@ -2795,7 +2847,7 @@ def format_float_or_int_string(text: str) -> str: def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None: """Make existing optional parentheses invisible or create new ones. - `parens_after` is a set of string leaf values immeditely after which parens + `parens_after` is a set of string leaf values immediately after which parens should be put. Standardizes on visible parentheses for single-element tuples, and keeps @@ -2817,6 +2869,8 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None: check_lpar = True if check_lpar: + if is_walrus_assignment(child): + continue if child.type == syms.atom: if maybe_make_parens_invisible_in_atom(child, parent=node): lpar = Leaf(token.LPAR, "") @@ -2971,18 +3025,24 @@ def is_empty_tuple(node: LN) -> bool: ) +def unwrap_singleton_parenthesis(node: LN) -> Optional[LN]: + """Returns `wrapped` if `node` is of the shape ( wrapped ). + + Parenthesis can be optional. Returns None otherwise""" + if len(node.children) != 3: + return None + lpar, wrapped, rpar = node.children + if not (lpar.type == token.LPAR and rpar.type == token.RPAR): + return None + + return wrapped + + def is_one_tuple(node: LN) -> bool: """Return True if `node` holds a tuple with one element, with or without parens.""" if node.type == syms.atom: - if len(node.children) != 3: - return False - - lpar, gexp, rpar = node.children - if not ( - lpar.type == token.LPAR - and gexp.type == syms.testlist_gexp - and rpar.type == token.RPAR - ): + gexp = unwrap_singleton_parenthesis(node) + if gexp is None or gexp.type != syms.testlist_gexp: return False return len(gexp.children) == 2 and gexp.children[1].type == token.COMMA @@ -2994,6 +3054,12 @@ def is_one_tuple(node: LN) -> bool: ) +def is_walrus_assignment(node: LN) -> bool: + """Return True iff `node` is of the shape ( test := test )""" + inner = unwrap_singleton_parenthesis(node) + return inner is not None and inner.type == syms.namedexpr_test + + def is_yield(node: LN) -> bool: """Return True if `node` holds a `yield` or `yield from` expression.""" if node.type == syms.yield_expr: @@ -3023,7 +3089,7 @@ def is_vararg(leaf: Leaf, within: Set[NodeType]) -> bool: extended iterable unpacking (PEP 3132) and additional unpacking generalizations (PEP 448). """ - if leaf.type not in STARS or not leaf.parent: + if leaf.type not in VARARGS_SPECIALS or not leaf.parent: return False p = leaf.parent @@ -3073,7 +3139,7 @@ def is_stub_body(node: LN) -> bool: ) -def max_delimiter_priority_in_atom(node: LN) -> int: +def max_delimiter_priority_in_atom(node: LN) -> Priority: """Return maximum delimiter priority inside `node`. This is specific to atoms with contents contained in a pair of parentheses. @@ -3105,7 +3171,7 @@ def ensure_visible(leaf: Leaf) -> None: """Make sure parentheses are visible. They could be invisible as part of some statements (see - :func:`normalize_invible_parens` and :func:`visit_import_from`). + :func:`normalize_invisible_parens` and :func:`visit_import_from`). """ if leaf.type == token.LPAR: leaf.value = "(" @@ -3138,8 +3204,9 @@ def get_features_used(node: Node) -> Set[Feature]: Currently looking for: - f-strings; - - underscores in numeric literals; and - - trailing commas after * or ** in function signatures and calls. + - underscores in numeric literals; + - trailing commas after * or ** in function signatures and calls; + - positional only arguments in function signatures and lambdas; """ features: Set[Feature] = set() for n in node.pre_order(): @@ -3152,6 +3219,13 @@ def get_features_used(node: Node) -> Set[Feature]: if "_" in n.value: # type: ignore features.add(Feature.NUMERIC_UNDERSCORES) + elif n.type == token.SLASH: + if n.parent and n.parent.type in {syms.typedargslist, syms.arglist}: + features.add(Feature.POS_ONLY_ARGUMENTS) + + elif n.type == token.COLONEQUAL: + features.add(Feature.ASSIGNMENT_EXPRESSIONS) + elif ( n.type in {syms.typedargslist, syms.arglist} and n.children @@ -3433,34 +3507,58 @@ class Report: return ", ".join(report) + "." -def parse_ast(src: str) -> Union[ast3.AST, ast27.AST]: - for feature_version in (7, 6): - try: - return ast3.parse(src, feature_version=feature_version) - except SyntaxError: - continue +def parse_ast(src: str) -> Union[ast.AST, ast3.AST, ast27.AST]: + filename = "" + if sys.version_info >= (3, 8): + # TODO: support Python 4+ ;) + for minor_version in range(sys.version_info[1], 4, -1): + try: + return ast.parse(src, filename, feature_version=(3, minor_version)) + except SyntaxError: + continue + else: + for feature_version in (7, 6): + try: + return ast3.parse(src, filename, feature_version=feature_version) + except SyntaxError: + continue return ast27.parse(src) +def _fixup_ast_constants( + node: Union[ast.AST, ast3.AST, ast27.AST] +) -> Union[ast.AST, ast3.AST, ast27.AST]: + """Map ast nodes deprecated in 3.8 to Constant.""" + # casts are required until this is released: + # https://github.com/python/typeshed/pull/3142 + if isinstance(node, (ast.Str, ast3.Str, ast27.Str, ast.Bytes, ast3.Bytes)): + return cast(ast.AST, ast.Constant(value=node.s)) + elif isinstance(node, (ast.Num, ast3.Num, ast27.Num)): + return cast(ast.AST, ast.Constant(value=node.n)) + elif isinstance(node, (ast.NameConstant, ast3.NameConstant)): + return cast(ast.AST, ast.Constant(value=node.value)) + return node + + def assert_equivalent(src: str, dst: str) -> None: """Raise AssertionError if `src` and `dst` aren't equivalent.""" - import traceback - - def _v(node: Union[ast3.AST, ast27.AST], depth: int = 0) -> Iterator[str]: + def _v(node: Union[ast.AST, ast3.AST, ast27.AST], depth: int = 0) -> Iterator[str]: """Simple visitor generating strings to compare ASTs by content.""" + + node = _fixup_ast_constants(node) + yield f"{' ' * depth}{node.__class__.__name__}(" for field in sorted(node._fields): # TypeIgnore has only one field 'lineno' which breaks this comparison - if isinstance(node, (ast3.TypeIgnore, ast27.TypeIgnore)): + type_ignore_classes = (ast3.TypeIgnore, ast27.TypeIgnore) + if sys.version_info >= (3, 8): + type_ignore_classes += (ast.TypeIgnore,) + if isinstance(node, type_ignore_classes): break - # Ignore str kind which is case sensitive / and ignores unicode_literals - if isinstance(node, (ast3.Str, ast27.Str, ast3.Bytes)) and field == "kind": - continue - try: value = getattr(node, field) except AttributeError: @@ -3474,15 +3572,15 @@ def assert_equivalent(src: str, dst: str) -> None: # parentheses and they change the AST. if ( field == "targets" - and isinstance(node, (ast3.Delete, ast27.Delete)) - and isinstance(item, (ast3.Tuple, ast27.Tuple)) + and isinstance(node, (ast.Delete, ast3.Delete, ast27.Delete)) + and isinstance(item, (ast.Tuple, ast3.Tuple, ast27.Tuple)) ): for item in item.elts: yield from _v(item, depth + 2) - elif isinstance(item, (ast3.AST, ast27.AST)): + elif isinstance(item, (ast.AST, ast3.AST, ast27.AST)): yield from _v(item, depth + 2) - elif isinstance(value, (ast3.AST, ast27.AST)): + elif isinstance(value, (ast.AST, ast3.AST, ast27.AST)): yield from _v(value, depth + 2) else: @@ -3504,7 +3602,7 @@ def assert_equivalent(src: str, dst: str) -> None: log = dump_to_file("".join(traceback.format_tb(exc.__traceback__)), dst) raise AssertionError( f"INTERNAL ERROR: Black produced invalid code: {exc}. " - f"Please report a bug on https://github.com/python/black/issues. " + f"Please report a bug on https://github.com/psf/black/issues. " f"This invalid output might be helpful: {log}" ) from None @@ -3515,7 +3613,7 @@ def assert_equivalent(src: str, dst: str) -> None: raise AssertionError( f"INTERNAL ERROR: Black produced code that is not equivalent to " f"the source. " - f"Please report a bug on https://github.com/python/black/issues. " + f"Please report a bug on https://github.com/psf/black/issues. " f"This diff might be helpful: {log}" ) from None @@ -3531,15 +3629,13 @@ def assert_stable(src: str, dst: str, mode: FileMode) -> None: raise AssertionError( f"INTERNAL ERROR: Black produced different code on the second pass " f"of the formatter. " - f"Please report a bug on https://github.com/python/black/issues. " + f"Please report a bug on https://github.com/psf/black/issues. " f"This diff might be helpful: {log}" ) from None def dump_to_file(*output: str) -> str: """Dump `output` to a temporary file. Return path to the file.""" - import tempfile - with tempfile.NamedTemporaryFile( mode="w", prefix="blk_", suffix=".log", delete=False, encoding="utf8" ) as f: @@ -3550,6 +3646,13 @@ def dump_to_file(*output: str) -> str: return f.name +@contextmanager +def nullcontext() -> Iterator[None]: + """Return context manager that does nothing. + Similar to `nullcontext` from python 3.7""" + yield + + def diff(a: str, b: str, a_name: str, b_name: str) -> str: """Return a unified diff string between strings `a` and `b`.""" import difflib @@ -3568,7 +3671,7 @@ def cancel(tasks: Iterable[asyncio.Task]) -> None: task.cancel() -def shutdown(loop: BaseEventLoop) -> None: +def shutdown(loop: asyncio.AbstractEventLoop) -> None: """Cancel all pending tasks on `loop`, wait for them, and close the loop.""" try: if sys.version_info[:2] >= (3, 7): @@ -3637,7 +3740,6 @@ def enumerate_with_length( if "\n" in leaf.value: return # Multiline strings, we can't continue. - comment: Optional[Leaf] for comment in line.comments_after(leaf): length += len(comment.value)