X-Git-Url: https://git.madduck.net/etc/vim.git/blobdiff_plain/180bc06ede96959fd52b43467ce3836655108719..3272bbc8b3e17d229df63daf26d3dce9706cee77:/black.py diff --git a/black.py b/black.py index 6dda93b..4599bdd 100644 --- a/black.py +++ b/black.py @@ -2,7 +2,7 @@ import asyncio import pickle from asyncio.base_events import BaseEventLoop from concurrent.futures import Executor, ProcessPoolExecutor -from enum import Enum +from enum import Enum, Flag from functools import partial, wraps import keyword import logging @@ -44,11 +44,12 @@ from blib2to3.pgen2 import driver, token from blib2to3.pgen2.parse import ParseError -__version__ = "18.4a6" +__version__ = "18.5b1" DEFAULT_LINE_LENGTH = 88 +CACHE_DIR = Path(user_cache_dir("black", version=__version__)) + # types -syms = pygram.python_symbols FileContent = str Encoding = str Depth = int @@ -65,6 +66,9 @@ Cache = Dict[Path, CacheInfo] out = partial(click.secho, bold=True, err=True) err = partial(click.secho, fg="red", err=True) +pygram.initialize(CACHE_DIR) +syms = pygram.python_symbols + class NothingChanged(UserWarning): """Raised by :func:`format_file` when reformatted code is the same as source.""" @@ -118,6 +122,13 @@ class Changed(Enum): YES = 2 +class FileMode(Flag): + AUTO_DETECT = 0 + PYTHON36 = 1 + PYI = 2 + NO_STRING_NORMALIZATION = 4 + + @click.command() @click.option( "-l", @@ -155,6 +166,29 @@ class Changed(Enum): "silence those with 2>/dev/null." ), ) +@click.option( + "--pyi", + is_flag=True, + help=( + "Consider all input files typing stubs regardless of file extension " + "(useful when piping source on standard input)." + ), +) +@click.option( + "--py36", + is_flag=True, + help=( + "Allow using Python 3.6-only syntax on all input files. This will put " + "trailing commas in function signatures and calls also after *args and " + "**kwargs. [default: per-file auto-detection]" + ), +) +@click.option( + "-S", + "--skip-string-normalization", + is_flag=True, + help="Don't normalize string quotes or prefixes.", +) @click.version_option(version=__version__) @click.argument( "src", @@ -170,6 +204,9 @@ def main( check: bool, diff: bool, fast: bool, + pyi: bool, + py36: bool, + skip_string_normalization: bool, quiet: bool, src: List[str], ) -> None: @@ -193,6 +230,13 @@ def main( write_back = WriteBack.DIFF else: write_back = WriteBack.YES + mode = FileMode.AUTO_DETECT + if py36: + mode |= FileMode.PYTHON36 + if pyi: + mode |= FileMode.PYI + if skip_string_normalization: + mode |= FileMode.NO_STRING_NORMALIZATION report = Report(check=check, quiet=quiet) if len(sources) == 0: out("No paths given. Nothing to do 😴") @@ -200,14 +244,28 @@ def main( return elif len(sources) == 1: - reformat_one(sources[0], line_length, fast, write_back, report) + reformat_one( + src=sources[0], + line_length=line_length, + fast=fast, + write_back=write_back, + mode=mode, + report=report, + ) else: loop = asyncio.get_event_loop() executor = ProcessPoolExecutor(max_workers=os.cpu_count()) try: loop.run_until_complete( schedule_formatting( - sources, line_length, fast, write_back, report, loop, executor + sources=sources, + line_length=line_length, + fast=fast, + write_back=write_back, + mode=mode, + report=report, + loop=loop, + executor=executor, ) ) finally: @@ -219,33 +277,43 @@ def main( def reformat_one( - src: Path, line_length: int, fast: bool, write_back: WriteBack, report: "Report" + src: Path, + line_length: int, + fast: bool, + write_back: WriteBack, + mode: FileMode, + report: "Report", ) -> 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`, and `fast` options are passed to :func:`format_file_in_place`. + `write_back`, `fast` and `pyi` options are passed to + :func:`format_file_in_place` or :func:`format_stdin_to_stdout`. """ try: changed = Changed.NO if not src.is_file() and str(src) == "-": if format_stdin_to_stdout( - line_length=line_length, fast=fast, write_back=write_back + line_length=line_length, fast=fast, write_back=write_back, mode=mode ): changed = Changed.YES else: cache: Cache = {} if write_back != WriteBack.DIFF: - cache = read_cache(line_length) + cache = read_cache(line_length, mode) 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 + src, + line_length=line_length, + fast=fast, + write_back=write_back, + mode=mode, ): changed = Changed.YES if write_back == WriteBack.YES and changed is not Changed.NO: - write_cache(cache, [src], line_length) + write_cache(cache, [src], line_length, mode) report.done(src, changed) except Exception as exc: report.failed(src, str(exc)) @@ -256,6 +324,7 @@ async def schedule_formatting( line_length: int, fast: bool, write_back: WriteBack, + mode: FileMode, report: "Report", loop: BaseEventLoop, executor: Executor, @@ -264,12 +333,12 @@ async def schedule_formatting( (Use ProcessPoolExecutors for actual parallelism.) - `line_length`, `write_back`, and `fast` options are passed to + `line_length`, `write_back`, `fast`, and `pyi` options are passed to :func:`format_file_in_place`. """ cache: Cache = {} if write_back != WriteBack.DIFF: - cache = read_cache(line_length) + cache = read_cache(line_length, mode) sources, cached = filter_cached(cache, sources) for src in cached: report.done(src, Changed.CACHED) @@ -284,7 +353,14 @@ async def schedule_formatting( lock = manager.Lock() tasks = { loop.run_in_executor( - executor, format_file_in_place, src, line_length, fast, write_back, lock + executor, + format_file_in_place, + src, + line_length, + fast, + write_back, + mode, + lock, ): src for src in sorted(sources) } @@ -309,7 +385,7 @@ async def schedule_formatting( if cancelled: await asyncio.gather(*cancelled, loop=loop, return_exceptions=True) if write_back == WriteBack.YES and formatted: - write_cache(cache, formatted, line_length) + write_cache(cache, formatted, line_length, mode) def format_file_in_place( @@ -317,6 +393,7 @@ def format_file_in_place( line_length: int, fast: bool, write_back: WriteBack = WriteBack.NO, + mode: FileMode = FileMode.AUTO_DETECT, lock: Any = None, # multiprocessing.Manager().Lock() is some crazy proxy ) -> bool: """Format file under `src` path. Return True if changed. @@ -324,13 +401,13 @@ def format_file_in_place( If `write_back` is True, write reformatted code back to stdout. `line_length` and `fast` options are passed to :func:`format_file_contents`. """ - is_pyi = src.suffix == ".pyi" - + if src.suffix == ".pyi": + mode |= FileMode.PYI with tokenize.open(src) as src_buffer: src_contents = src_buffer.read() try: dst_contents = format_file_contents( - src_contents, line_length=line_length, fast=fast, is_pyi=is_pyi + src_contents, line_length=line_length, fast=fast, mode=mode ) except NothingChanged: return False @@ -353,17 +430,21 @@ def format_file_in_place( def format_stdin_to_stdout( - line_length: int, fast: bool, write_back: WriteBack = WriteBack.NO + line_length: int, + fast: bool, + write_back: WriteBack = WriteBack.NO, + mode: FileMode = FileMode.AUTO_DETECT, ) -> bool: """Format file on stdin. Return True if changed. If `write_back` is True, write reformatted code back to stdout. - `line_length` and `fast` arguments are passed to :func:`format_file_contents`. + `line_length`, `fast`, `is_pyi`, and `force_py36` arguments are passed to + :func:`format_file_contents`. """ src = sys.stdin.read() dst = src try: - dst = format_file_contents(src, line_length=line_length, fast=fast) + dst = format_file_contents(src, line_length=line_length, fast=fast, mode=mode) return True except NothingChanged: @@ -379,7 +460,11 @@ def format_stdin_to_stdout( def format_file_contents( - src_contents: str, *, line_length: int, fast: bool, is_pyi: bool = False + src_contents: str, + *, + line_length: int, + fast: bool, + mode: FileMode = FileMode.AUTO_DETECT, ) -> FileContent: """Reformat contents a file and return new contents. @@ -390,20 +475,18 @@ def format_file_contents( if src_contents.strip() == "": raise NothingChanged - dst_contents = format_str(src_contents, line_length=line_length, is_pyi=is_pyi) + dst_contents = format_str(src_contents, line_length=line_length, mode=mode) if src_contents == dst_contents: raise NothingChanged if not fast: assert_equivalent(src_contents, dst_contents) - assert_stable( - src_contents, dst_contents, line_length=line_length, is_pyi=is_pyi - ) + assert_stable(src_contents, dst_contents, line_length=line_length, mode=mode) return dst_contents def format_str( - src_contents: str, line_length: int, *, is_pyi: bool = False + src_contents: str, line_length: int, *, mode: FileMode = FileMode.AUTO_DETECT ) -> FileContent: """Reformat a string and return new contents. @@ -412,11 +495,15 @@ def format_str( src_node = lib2to3_parse(src_contents) dst_contents = "" future_imports = get_future_imports(src_node) - elt = EmptyLineTracker(is_pyi=is_pyi) - py36 = is_python36(src_node) + is_pyi = bool(mode & FileMode.PYI) + py36 = bool(mode & FileMode.PYTHON36) or is_python36(src_node) + normalize_strings = not bool(mode & FileMode.NO_STRING_NORMALIZATION) lines = LineGenerator( - remove_u_prefix=py36 or "unicode_literals" in future_imports, is_pyi=is_pyi + remove_u_prefix=py36 or "unicode_literals" in future_imports, + is_pyi=is_pyi, + normalize_strings=normalize_strings, ) + elt = EmptyLineTracker(is_pyi=is_pyi) empty_line = Line() after = 0 for current_line in lines.visit(src_node): @@ -874,27 +961,6 @@ class Line: and second_leaf.value == "def" ) - @property - def is_flow_control(self) -> bool: - """Is this line a flow control statement? - - Those are `return`, `raise`, `break`, and `continue`. - """ - return ( - bool(self) - and self.leaves[0].type == token.NAME - and self.leaves[0].value in FLOW_CONTROL - ) - - @property - def is_yield(self) -> bool: - """Is this line a yield statement?""" - return ( - bool(self) - and self.leaves[0].type == token.NAME - and self.leaves[0].value == "yield" - ) - @property def is_class_paren_empty(self) -> bool: """Is this a class with no base classes but using parentheses? @@ -911,6 +977,15 @@ class Line: and self.leaves[3].value == ")" ) + @property + def is_triple_quoted_string(self) -> bool: + """Is the line a triple quoted string?""" + return ( + bool(self) + and self.leaves[0].type == token.STRING + and self.leaves[0].value.startswith(('"""', "'''")) + ) + def contains_standalone_comments(self, depth_limit: int = sys.maxsize) -> bool: """If so, needs to be split before emitting.""" for leaf in self.leaves: @@ -1118,6 +1193,7 @@ class EmptyLineTracker: the prefix of the first leaf consists of optional newlines. Those newlines are consumed by `maybe_empty_lines()` and included in the computation. """ + is_pyi: bool = False previous_line: Optional[Line] = None previous_after: int = 0 @@ -1127,8 +1203,7 @@ class EmptyLineTracker: """Return the number of extra empty lines before and after the `current_line`. This is for separating `def`, `async def` and `class` with extra empty - lines (two on module-level), as well as providing an extra empty line - after flow control keywords to make them more prominent. + lines (two on module-level). """ if isinstance(current_line, UnformattedLines): return 0, 0 @@ -1169,6 +1244,11 @@ class EmptyLineTracker: if self.previous_line.is_decorator: return 0, 0 + if self.previous_line.depth < current_line.depth and ( + self.previous_line.is_class or self.previous_line.is_def + ): + return 0, 0 + if ( self.previous_line.is_comment and self.previous_line.depth == current_line.depth @@ -1200,6 +1280,13 @@ class EmptyLineTracker: ): return (before or 1), 0 + if ( + self.previous_line + and self.previous_line.is_class + and current_line.is_triple_quoted_string + ): + return before, 1 + return before, 0 @@ -1210,7 +1297,9 @@ class LineGenerator(Visitor[Line]): Note: destroys the tree it's visiting by mutating prefixes of its leaves in ways that will no longer stringify to valid Python code on the tree. """ + is_pyi: bool = False + normalize_strings: bool = True current_line: Line = Factory(Line) remove_u_prefix: bool = False @@ -1279,7 +1368,7 @@ class LineGenerator(Visitor[Line]): else: normalize_prefix(node, inside_brackets=any_open_brackets) - if node.type == token.STRING: + if self.normalize_strings and node.type == token.STRING: normalize_string_prefix(node, remove_u_prefix=self.remove_u_prefix) normalize_string_quotes(node) if node.type not in WHITESPACE: @@ -1371,32 +1460,6 @@ class LineGenerator(Visitor[Line]): yield from self.line() yield from self.visit(child) - def visit_import_from(self, node: Node) -> Iterator[Line]: - """Visit import_from and maybe put invisible parentheses. - - This is separate from `visit_stmt` because import statements don't - support arbitrary atoms and thus handling of parentheses is custom. - """ - check_lpar = False - for index, child in enumerate(node.children): - if check_lpar: - if child.type == token.LPAR: - # make parentheses invisible - child.value = "" # type: ignore - node.children[-1].value = "" # type: ignore - else: - # insert invisible parentheses - node.insert_child(index, Leaf(token.LPAR, "")) - node.append_child(Leaf(token.RPAR, "")) - break - - check_lpar = ( - child.type == token.NAME and child.value == "import" # type: ignore - ) - - for child in node.children: - yield from self.visit(child) - def visit_SEMI(self, leaf: Leaf) -> Iterator[Line]: """Remove a semicolon and put the other statement on a separate line.""" yield from self.line() @@ -1443,6 +1506,7 @@ class LineGenerator(Visitor[Line]): self.visit_classdef = partial(v, keywords={"class"}, parens=Ø) self.visit_expr_stmt = partial(v, keywords=Ø, parens=ASSIGNMENTS) self.visit_return_stmt = partial(v, keywords={"return"}, parens={"return"}) + self.visit_import_from = partial(v, keywords=Ø, parens={"import"}) self.visit_async_funcdef = self.visit_async_stmt self.visit_decorated = self.visit_decorators @@ -1816,7 +1880,7 @@ def is_split_before_delimiter(leaf: Leaf, previous: Leaf = None) -> int: return 0 -def generate_comments(leaf: Leaf) -> Iterator[Leaf]: +def generate_comments(leaf: LN) -> Iterator[Leaf]: """Clean the prefix of the `leaf` and generate comments from it, if any. Comments in lib2to3 are shoved into the whitespace prefix. This happens @@ -2333,8 +2397,13 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None: Standardizes on visible parentheses for single-element tuples, and keeps existing visible parentheses for other tuples and generator expressions. """ + try: + list(generate_comments(node)) + except FormatOff: + return # This `node` has a prefix with `# fmt: off`, don't mess with parens. + check_lpar = False - for child in list(node.children): + for index, child in enumerate(list(node.children)): if check_lpar: if child.type == syms.atom: maybe_make_parens_invisible_in_atom(child) @@ -2342,8 +2411,21 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None: # wrap child in visible parentheses lpar = Leaf(token.LPAR, "(") rpar = Leaf(token.RPAR, ")") - index = child.remove() or 0 + child.remove() node.insert_child(index, Node(syms.atom, [lpar, child, rpar])) + elif node.type == syms.import_from: + # "import from" nodes store parentheses directly as part of + # the statement + if child.type == token.LPAR: + # make parentheses invisible + child.value = "" # type: ignore + node.children[-1].value = "" # type: ignore + elif child.type != token.STAR: + # insert invisible parentheses + node.insert_child(index, Leaf(token.LPAR, "")) + node.append_child(Leaf(token.RPAR, "")) + break + elif not (isinstance(child, Leaf) and is_multiline_string(child)): # wrap child in invisible parentheses lpar = Leaf(token.LPAR, "") @@ -2602,6 +2684,10 @@ def generate_trailers_to_omit(line: Line, line_length: int) -> Iterator[Set[Leaf if length > line_length: break + has_inline_comment = leaf_length > len(leaf.value) + len(leaf.prefix) + if leaf.type == STANDALONE_COMMENT or has_inline_comment: + break + optional_brackets.discard(id(leaf)) if opening_bracket: if leaf is opening_bracket: @@ -2696,6 +2782,7 @@ def gen_python_files_in_dir(path: Path) -> Iterator[Path]: @dataclass class Report: """Provides a reformatting counter. Can be rendered with `str(report)`.""" + check: bool = False quiet: bool = False change_count: int = 0 @@ -2835,9 +2922,11 @@ def assert_equivalent(src: str, dst: str) -> None: ) from None -def assert_stable(src: str, dst: str, line_length: int, is_pyi: bool = False) -> None: +def assert_stable( + src: str, dst: str, line_length: int, mode: FileMode = FileMode.AUTO_DETECT +) -> None: """Raise AssertionError if `dst` reformats differently the second time.""" - newdst = format_str(dst, line_length=line_length, is_pyi=is_pyi) + newdst = format_str(dst, line_length=line_length, mode=mode) if dst != newdst: log = dump_to_file( diff(src, dst, "source", "first pass"), @@ -2940,9 +3029,6 @@ def enumerate_with_length( comment: Optional[Leaf] for comment in line.comments_after(leaf, index): - if "\n" in comment.prefix: - return # Oops, standalone comment! - length += len(comment.value) yield index, leaf, length @@ -3051,19 +3137,21 @@ def can_omit_invisible_parens(line: Line, line_length: int) -> bool: return False -CACHE_DIR = Path(user_cache_dir("black", version=__version__)) - - -def get_cache_file(line_length: int) -> Path: - return CACHE_DIR / f"cache.{line_length}.pickle" +def get_cache_file(line_length: int, mode: FileMode) -> Path: + pyi = bool(mode & FileMode.PYI) + py36 = bool(mode & FileMode.PYTHON36) + return ( + CACHE_DIR + / f"cache.{line_length}{'.pyi' if pyi else ''}{'.py36' if py36 else ''}.pickle" + ) -def read_cache(line_length: int) -> Cache: +def read_cache(line_length: int, mode: FileMode) -> Cache: """Read the cache if it exists and is well formed. If it is not well formed, the call to write_cache later should resolve the issue. """ - cache_file = get_cache_file(line_length) + cache_file = get_cache_file(line_length, mode) if not cache_file.exists(): return {} @@ -3100,9 +3188,11 @@ def filter_cached( return todo, done -def write_cache(cache: Cache, sources: List[Path], line_length: int) -> None: +def write_cache( + cache: Cache, sources: List[Path], line_length: int, mode: FileMode +) -> None: """Update the cache file.""" - cache_file = get_cache_file(line_length) + cache_file = get_cache_file(line_length, mode) try: if not CACHE_DIR.exists(): CACHE_DIR.mkdir(parents=True)