X-Git-Url: https://git.madduck.net/etc/vim.git/blobdiff_plain/14ba1bf8b6248e6860ba6a0cb9468c4c1c25a102..31fbd1982f9511c5b7bce93eb133110811f90e0f:/black.py?ds=sidebyside diff --git a/black.py b/black.py index 81241f6..a1a3468 100644 --- a/black.py +++ b/black.py @@ -239,11 +239,8 @@ def reformat_one( src = src.resolve() if src in cache and cache[src] == get_cache_info(src): changed = Changed.CACHED - if ( - changed is not Changed.CACHED - and format_file_in_place( - src, line_length=line_length, fast=fast, write_back=write_back - ) + if changed is not Changed.CACHED and format_file_in_place( + src, line_length=line_length, fast=fast, write_back=write_back ): changed = Changed.YES if write_back == WriteBack.YES and changed is not Changed.NO: @@ -285,32 +282,29 @@ async def schedule_formatting( manager = Manager() lock = manager.Lock() tasks = { - src: loop.run_in_executor( + loop.run_in_executor( executor, format_file_in_place, src, line_length, fast, write_back, lock - ) - for src in sources + ): src + for src in sorted(sources) } - _task_values = list(tasks.values()) + pending: Iterable[asyncio.Task] = tasks.keys() try: - loop.add_signal_handler(signal.SIGINT, cancel, _task_values) - loop.add_signal_handler(signal.SIGTERM, cancel, _task_values) + loop.add_signal_handler(signal.SIGINT, cancel, pending) + loop.add_signal_handler(signal.SIGTERM, cancel, pending) except NotImplementedError: # There are no good alternatives for these on Windows pass - await asyncio.wait(_task_values) - for src, task in tasks.items(): - if not task.done(): - report.failed(src, "timed out, cancelling") - task.cancel() - cancelled.append(task) - elif task.cancelled(): - cancelled.append(task) - elif task.exception(): - report.failed(src, str(task.exception())) - else: - formatted.append(src) - report.done(src, Changed.YES if task.result() else Changed.NO) - + while pending: + done, _ = await asyncio.wait(pending, return_when=asyncio.FIRST_COMPLETED) + for task in done: + src = tasks.pop(task) + if task.cancelled(): + cancelled.append(task) + elif task.exception(): + report.failed(src, str(task.exception())) + else: + formatted.append(src) + report.done(src, Changed.YES if task.result() else Changed.NO) if cancelled: await asyncio.gather(*cancelled, loop=loop, return_exceptions=True) if write_back == WriteBack.YES and formatted: @@ -632,21 +626,22 @@ LOGIC_PRIORITY = 14 STRING_PRIORITY = 12 COMPARATOR_PRIORITY = 10 MATH_PRIORITIES = { - token.VBAR: 8, - token.CIRCUMFLEX: 7, - token.AMPER: 6, - token.LEFTSHIFT: 5, - token.RIGHTSHIFT: 5, - token.PLUS: 4, - token.MINUS: 4, - token.STAR: 3, - token.SLASH: 3, - token.DOUBLESLASH: 3, - token.PERCENT: 3, - token.AT: 3, - token.TILDE: 2, - token.DOUBLESTAR: 1, + token.VBAR: 9, + token.CIRCUMFLEX: 8, + token.AMPER: 7, + token.LEFTSHIFT: 6, + token.RIGHTSHIFT: 6, + token.PLUS: 5, + token.MINUS: 5, + token.STAR: 4, + token.SLASH: 4, + token.DOUBLESLASH: 4, + token.PERCENT: 4, + token.AT: 4, + token.TILDE: 3, + token.DOUBLESTAR: 2, } +DOT_PRIORITY = 1 @dataclass @@ -712,6 +707,17 @@ 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: + """Return the number of delimiters with the given `priority`. + + If no `priority` is passed, defaults to max priority on the line. + """ + if not self.delimiters: + return 0 + + priority = priority or self.max_delimiter_priority() + return sum(1 for p in self.delimiters.values() if p == priority) + def maybe_increment_for_loop_variable(self, leaf: Leaf) -> bool: """In a for loop, or comprehension, the variables are often unpacks. @@ -770,6 +776,7 @@ class Line: comments: List[Tuple[Index, Leaf]] = Factory(list) bracket_tracker: BracketTracker = Factory(BracketTracker) inside_brackets: bool = False + should_explode: bool = False def append(self, leaf: Leaf, preformatted: bool = False) -> None: """Add a new `leaf` to the end of the line. @@ -841,12 +848,11 @@ class Line: ) @property - def is_trivial_class(self) -> bool: + def is_stub_class(self) -> bool: """Is this line a class definition with a body consisting only of "..."?""" - return ( - self.is_class - and self.leaves[-3:] == [Leaf(token.DOT, ".") for _ in range(3)] - ) + return self.is_class and self.leaves[-3:] == [ + Leaf(token.DOT, ".") for _ in range(3) + ] @property def is_def(self) -> bool: @@ -860,14 +866,11 @@ class Line: second_leaf: Optional[Leaf] = self.leaves[1] except IndexError: second_leaf = None - return ( - (first_leaf.type == token.NAME and first_leaf.value == "def") - or ( - first_leaf.type == token.ASYNC - and second_leaf is not None - and second_leaf.type == token.NAME - and second_leaf.value == "def" - ) + return (first_leaf.type == token.NAME and first_leaf.value == "def") or ( + first_leaf.type == token.ASYNC + and second_leaf is not None + and second_leaf.type == token.NAME + and second_leaf.value == "def" ) @property @@ -1032,9 +1035,8 @@ class Line: and subscript_start.type == syms.subscriptlist ): subscript_start = child_towards(subscript_start, leaf) - return ( - subscript_start is not None - and any(n.type in TEST_DESCENDANTS for n in subscript_start.pre_order()) + return subscript_start is not None and any( + n.type in TEST_DESCENDANTS for n in subscript_start.pre_order() ) def __str__(self) -> str: @@ -1177,10 +1179,7 @@ class EmptyLineTracker: if self.previous_line.depth > current_line.depth: newlines = 1 elif current_line.is_class or self.previous_line.is_class: - if ( - current_line.is_trivial_class - and self.previous_line.is_trivial_class - ): + if current_line.is_stub_class and self.previous_line.is_stub_class: newlines = 0 else: newlines = 1 @@ -1317,7 +1316,7 @@ class LineGenerator(Visitor[Line]): The relevant Python language `keywords` for a given statement will be NAME leaves within it. This methods puts those on a separate line. - `parens` holds a set of string leaf values immeditely after which + `parens` holds a set of string leaf values immediately after which invisible parens should be put. """ normalize_invisible_parens(node, parens_after=parens) @@ -1329,51 +1328,16 @@ class LineGenerator(Visitor[Line]): def visit_suite(self, node: Node) -> Iterator[Line]: """Visit a suite.""" - if self.is_pyi and self.is_trivial_suite(node): + if self.is_pyi and is_stub_suite(node): yield from self.visit(node.children[2]) else: yield from self.visit_default(node) - def is_trivial_suite(self, node: Node) -> bool: - if len(node.children) != 4: - return False - if ( - not isinstance(node.children[0], Leaf) - or node.children[0].type != token.NEWLINE - ): - return False - if ( - not isinstance(node.children[1], Leaf) - or node.children[1].type != token.INDENT - ): - return False - if ( - not isinstance(node.children[3], Leaf) - or node.children[3].type != token.DEDENT - ): - return False - stmt = node.children[2] - if not isinstance(stmt, Node): - return False - return self.is_trivial_body(stmt) - - def is_trivial_body(self, stmt: Node) -> bool: - if not isinstance(stmt, Node) or stmt.type != syms.simple_stmt: - return False - if len(stmt.children) != 2: - return False - child = stmt.children[0] - return ( - child.type == syms.atom - and len(child.children) == 3 - and all(leaf == Leaf(token.DOT, ".") for leaf in child.children) - ) - def visit_simple_stmt(self, node: Node) -> Iterator[Line]: """Visit a statement without nested statements.""" is_suite_like = node.parent and node.parent.type in STATEMENT if is_suite_like: - if self.is_pyi and self.is_trivial_body(node): + if self.is_pyi and is_stub_body(node): yield from self.visit_default(node) else: yield from self.line(+1) @@ -1381,11 +1345,7 @@ class LineGenerator(Visitor[Line]): yield from self.line(-1) else: - if ( - not self.is_pyi - or not node.parent - or not self.is_trivial_suite(node.parent) - ): + if not self.is_pyi or not node.parent or not is_stub_suite(node.parent): yield from self.line() yield from self.visit_default(node) @@ -1513,10 +1473,11 @@ def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str: # noqa C901 return DOUBLESPACE assert p is not None, f"INTERNAL ERROR: hand-made leaf without parent: {leaf!r}" - if ( - t == token.COLON - and p.type not in {syms.subscript, syms.subscriptlist, syms.sliceop} - ): + if t == token.COLON and p.type not in { + syms.subscript, + syms.subscriptlist, + syms.sliceop, + }: return NO prev = leaf.prev_sibling @@ -1537,7 +1498,10 @@ def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str: # noqa C901 if prevp.type == token.EQUAL: if prevp.parent: if prevp.parent.type in { - syms.arglist, syms.argument, syms.parameters, syms.varargslist + syms.arglist, + syms.argument, + syms.parameters, + syms.varargslist, }: return NO @@ -1690,10 +1654,10 @@ def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str: # noqa C901 prevp_parent = prevp.parent assert prevp_parent is not None - if ( - prevp.type == token.COLON - and prevp_parent.type in {syms.subscript, syms.sliceop} - ): + if prevp.type == token.COLON and prevp_parent.type in { + syms.subscript, + syms.sliceop, + }: return NO elif prevp.type == token.EQUAL and prevp_parent.type == syms.argument: @@ -1773,6 +1737,14 @@ def is_split_before_delimiter(leaf: Leaf, previous: Leaf = None) -> int: # Don't treat them as a delimiter. return 0 + if ( + leaf.type == token.DOT + and leaf.parent + and leaf.parent.type not in {syms.import_from, syms.dotted_name} + and (previous is None or previous.type in CLOSING_BRACKETS) + ): + return DOT_PRIORITY + if ( leaf.type in MATH_OPERATORS and leaf.parent @@ -1790,31 +1762,54 @@ def is_split_before_delimiter(leaf: Leaf, previous: Leaf = None) -> int: ): return STRING_PRIORITY + if leaf.type != token.NAME: + return 0 + if ( - leaf.type == token.NAME - and leaf.value == "for" + leaf.value == "for" and leaf.parent and leaf.parent.type in {syms.comp_for, syms.old_comp_for} ): return COMPREHENSION_PRIORITY if ( - leaf.type == token.NAME - and leaf.value == "if" + leaf.value == "if" and leaf.parent and leaf.parent.type in {syms.comp_if, syms.old_comp_if} ): return COMPREHENSION_PRIORITY + if leaf.value in {"if", "else"} and leaf.parent and leaf.parent.type == syms.test: + return TERNARY_PRIORITY + + if leaf.value == "is": + return COMPARATOR_PRIORITY + if ( - leaf.type == token.NAME - and leaf.value in {"if", "else"} + leaf.value == "in" and leaf.parent - and leaf.parent.type == syms.test + and leaf.parent.type in {syms.comp_op, syms.comparison} + and not ( + previous is not None + and previous.type == token.NAME + and previous.value == "not" + ) ): - return TERNARY_PRIORITY + return COMPARATOR_PRIORITY - if leaf.type == token.NAME and leaf.value in LOGIC_OPERATORS and leaf.parent: + if ( + leaf.value == "not" + and leaf.parent + and leaf.parent.type == syms.comp_op + and not ( + previous is not None + and previous.type == token.NAME + and previous.value == "is" + ) + ): + return COMPARATOR_PRIORITY + + if leaf.value in LOGIC_OPERATORS and leaf.parent: return LOGIC_PRIORITY return 0 @@ -1914,15 +1909,15 @@ def split_line( return line_str = str(line).strip("\n") - if is_line_short_enough(line, line_length=line_length, line_str=line_str): + if not line.should_explode and is_line_short_enough( + line, line_length=line_length, line_str=line_str + ): yield line return split_funcs: List[SplitFunc] if line.is_def: split_funcs = [left_hand_split] - elif line.is_import: - split_funcs = [explode_split] else: def rhs(line: Line, py36: bool = False) -> Iterator[Line]: @@ -2013,6 +2008,8 @@ def right_hand_split( If the split was by optional parentheses, attempt splitting without them, too. `omit` is a collection of closing bracket IDs that shouldn't be considered for this split. + + Note: running this function modifies `bracket_depth` on the leaves of `line`. """ head = Line(depth=line.depth) body = Line(depth=line.depth + 1, inside_brackets=True) @@ -2039,8 +2036,9 @@ def right_hand_split( # Since body is a new indent level, remove spurious leading whitespace. if body_leaves: normalize_prefix(body_leaves[0], inside_brackets=True) - elif not head_leaves: - # No `head` and no `body` means the split failed. `tail` has all content. + if not head_leaves: + # No `head` means the split failed. Either `tail` has all content or + # the matching `opening_bracket` wasn't available on `line` anymore. raise CannotSplit("No brackets found") # Build the new lines. @@ -2058,22 +2056,31 @@ def right_hand_split( # the closing bracket is an optional paren and closing_bracket.type == token.RPAR and not closing_bracket.value - # there are no delimiters or standalone comments in the body - and not body.bracket_tracker.delimiters + # there are no standalone comments in the body and not line.contains_standalone_comments(0) # and it's not an import (optional parens are the only thing we can split # on in this case; attempting a split without them is a waste of time) and not line.is_import ): omit = {id(closing_bracket), *omit} - try: - yield from right_hand_split(line, py36=py36, omit=omit) - return - except CannotSplit: - pass + delimiter_count = body.bracket_tracker.delimiter_count_with_priority() + if ( + delimiter_count == 0 + or delimiter_count == 1 + and ( + body.leaves[0].type in OPENING_BRACKETS + or body.leaves[-1].type in CLOSING_BRACKETS + ) + ): + try: + yield from right_hand_split(line, py36=py36, omit=omit) + return + except CannotSplit: + pass ensure_visible(opening_bracket) ensure_visible(closing_bracket) + body.should_explode = should_explode(body, opening_bracket) for result in (head, body, tail): if result: yield result @@ -2132,14 +2139,16 @@ def delimiter_split(line: Line, py36: bool = False) -> Iterator[Line]: except IndexError: raise CannotSplit("Line empty") - delimiters = line.bracket_tracker.delimiters + bt = line.bracket_tracker try: - delimiter_priority = line.bracket_tracker.max_delimiter_priority( - exclude={id(last_leaf)} - ) + delimiter_priority = bt.max_delimiter_priority(exclude={id(last_leaf)}) except ValueError: raise CannotSplit("No delimiters found") + if delimiter_priority == DOT_PRIORITY: + if bt.delimiter_count_with_priority(delimiter_priority) == 1: + raise CannotSplit("Splitting a single attribute from its owner looks wrong") + current_line = Line(depth=line.depth, inside_brackets=line.inside_brackets) lowest_depth = sys.maxsize trailing_comma_safe = True @@ -2162,12 +2171,11 @@ def delimiter_split(line: Line, py36: bool = False) -> Iterator[Line]: yield from append_to_line(comment_after) lowest_depth = min(lowest_depth, leaf.bracket_depth) - if ( - leaf.bracket_depth == lowest_depth - and is_vararg(leaf, within=VARARGS_PARENTS) + if leaf.bracket_depth == lowest_depth and is_vararg( + leaf, within=VARARGS_PARENTS ): trailing_comma_safe = trailing_comma_safe and py36 - leaf_priority = delimiters.get(id(leaf)) + leaf_priority = bt.delimiters.get(id(leaf)) if leaf_priority == delimiter_priority: yield current_line @@ -2212,26 +2220,6 @@ def standalone_comment_split(line: Line, py36: bool = False) -> Iterator[Line]: yield current_line -def explode_split( - line: Line, py36: bool = False, omit: Collection[LeafID] = () -) -> Iterator[Line]: - """Split by rightmost bracket and immediately split contents by a delimiter.""" - new_lines = list(right_hand_split(line, py36, omit)) - if len(new_lines) != 3: - yield from new_lines - return - - yield new_lines[0] - - try: - yield from delimiter_split(new_lines[1], py36) - - except CannotSplit: - yield new_lines[1] - - yield new_lines[2] - - def is_import(leaf: Leaf) -> bool: """Return True if the given leaf starts an import statement.""" p = leaf.parent @@ -2361,7 +2349,7 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None: rpar = Leaf(token.RPAR, ")") index = child.remove() or 0 node.insert_child(index, Node(syms.atom, [lpar, child, rpar])) - else: + elif not (isinstance(child, Leaf) and is_multiline_string(child)): # wrap child in invisible parentheses lpar = Leaf(token.LPAR, "") rpar = Leaf(token.RPAR, "") @@ -2472,6 +2460,41 @@ def is_vararg(leaf: Leaf, within: Set[NodeType]) -> bool: return p.type in within +def is_multiline_string(leaf: Leaf) -> bool: + """Return True if `leaf` is a multiline string that actually spans many lines.""" + value = leaf.value.lstrip("furbFURB") + return value[:3] in {'"""', "'''"} and "\n" in value + + +def is_stub_suite(node: Node) -> bool: + """Return True if `node` is a suite with a stub body.""" + if ( + len(node.children) != 4 + or node.children[0].type != token.NEWLINE + or node.children[1].type != token.INDENT + or node.children[3].type != token.DEDENT + ): + return False + + return is_stub_body(node.children[2]) + + +def is_stub_body(node: LN) -> bool: + """Return True if `node` is a simple statement containing an ellipsis.""" + if not isinstance(node, Node) or node.type != syms.simple_stmt: + return False + + if len(node.children) != 2: + return False + + child = node.children[0] + return ( + child.type == syms.atom + and len(child.children) == 3 + and all(leaf == Leaf(token.DOT, ".") for leaf in child.children) + ) + + def max_delimiter_priority_in_atom(node: LN) -> int: """Return maximum delimiter priority inside `node`. @@ -2512,6 +2535,17 @@ def ensure_visible(leaf: Leaf) -> None: leaf.value = ")" +def should_explode(line: Line, opening_bracket: Leaf) -> bool: + """Should `line` immediately be split with `delimiter_split()` after RHS?""" + return bool( + opening_bracket.parent + and opening_bracket.parent.type in {syms.atom, syms.import_from} + and opening_bracket.value in "[{(" + and line.bracket_tracker.delimiters + and line.bracket_tracker.max_delimiter_priority() == COMMA_PRIORITY + ) + + def is_python36(node: Node) -> bool: """Return True if the current file is using Python 3.6+ features. @@ -2640,7 +2674,15 @@ def get_future_imports(node: Node) -> Set[str]: PYTHON_EXTENSIONS = {".py", ".pyi"} BLACKLISTED_DIRECTORIES = { - "build", "buck-out", "dist", "_build", ".git", ".hg", ".mypy_cache", ".tox", ".venv" + "build", + "buck-out", + "dist", + "_build", + ".git", + ".hg", + ".mypy_cache", + ".tox", + ".venv", } @@ -2842,7 +2884,7 @@ def diff(a: str, b: str, a_name: str, b_name: str) -> str: ) -def cancel(tasks: List[asyncio.Task]) -> None: +def cancel(tasks: Iterable[asyncio.Task]) -> None: """asyncio signal handler that cancels all `tasks` and reports to stderr.""" err("Aborted!") for task in tasks: