X-Git-Url: https://git.madduck.net/etc/vim.git/blobdiff_plain/b2067aabbfa900366326ab7ab7d5a208059f5dab..d638d56e0e383caeeff048e76b02b50b29367c3d:/black.py?ds=sidebyside diff --git a/black.py b/black.py index 488f581..551d3c1 100644 --- a/black.py +++ b/black.py @@ -1,18 +1,20 @@ import asyncio -import pickle from asyncio.base_events import BaseEventLoop from concurrent.futures import Executor, ProcessPoolExecutor +from datetime import datetime from enum import Enum, Flag from functools import partial, wraps +import io import keyword import logging from multiprocessing import Manager import os from pathlib import Path +import pickle import re -import tokenize import signal import sys +import tokenize from typing import ( Any, Callable, @@ -46,12 +48,17 @@ from blib2to3.pgen2.parse import ParseError __version__ = "18.5b1" DEFAULT_LINE_LENGTH = 88 +DEFAULT_EXCLUDES = ( + r"/(\.git|\.hg|\.mypy_cache|\.tox|\.venv|_build|buck-out|build|dist)/" +) +DEFAULT_INCLUDES = r"\.pyi?$" CACHE_DIR = Path(user_cache_dir("black", version=__version__)) # types FileContent = str Encoding = str +NewLine = str Depth = int NodeType = int LeafID = int @@ -115,6 +122,13 @@ class WriteBack(Enum): YES = 1 DIFF = 2 + @classmethod + def from_configuration(cls, *, check: bool, diff: bool) -> "WriteBack": + if check and not diff: + return cls.NO + + return cls.DIFF if diff else cls.YES + class Changed(Enum): NO = 0 @@ -126,6 +140,20 @@ class FileMode(Flag): AUTO_DETECT = 0 PYTHON36 = 1 PYI = 2 + NO_STRING_NORMALIZATION = 4 + + @classmethod + def from_configuration( + cls, *, py36: bool, pyi: bool, skip_string_normalization: bool + ) -> "FileMode": + mode = cls.AUTO_DETECT + if py36: + mode |= cls.PYTHON36 + if pyi: + mode |= cls.PYI + if skip_string_normalization: + mode |= cls.NO_STRING_NORMALIZATION + return mode @click.command() @@ -137,6 +165,29 @@ class FileMode(Flag): help="How many character per line to allow.", show_default=True, ) +@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( + "--pyi", + is_flag=True, + help=( + "Format all input files like typing stubs regardless of file extension " + "(useful when piping source on standard input)." + ), +) +@click.option( + "-S", + "--skip-string-normalization", + is_flag=True, + help="Don't normalize string quotes or prefixes.", +) @click.option( "--check", is_flag=True, @@ -157,29 +208,46 @@ class FileMode(Flag): help="If --fast given, skip temporary sanity checks. [default: --safe]", ) @click.option( - "-q", - "--quiet", - is_flag=True, + "--include", + type=str, + default=DEFAULT_INCLUDES, help=( - "Don't emit non-error messages to stderr. Errors are still emitted, " - "silence those with 2>/dev/null." + "A regular expression that matches files and directories that should be " + "included on recursive searches. An empty value means all files are " + "included regardless of the name. Use forward slashes for directories on " + "all platforms (Windows, too). Exclusions are calculated first, inclusions " + "later." ), + show_default=True, ) @click.option( - "--pyi", + "--exclude", + type=str, + default=DEFAULT_EXCLUDES, + help=( + "A regular expression that matches files and directories that should be " + "excluded on recursive searches. An empty value means no paths are excluded. " + "Use forward slashes for directories on all platforms (Windows, too). " + "Exclusions are calculated first, inclusions later." + ), + show_default=True, +) +@click.option( + "-q", + "--quiet", is_flag=True, help=( - "Consider all input files typing stubs regardless of file extension " - "(useful when piping source on standard input)." + "Don't emit non-error messages to stderr. Errors are still emitted, " + "silence those with 2>/dev/null." ), ) @click.option( - "--py36", + "-v", + "--verbose", 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]" + "Also emit messages to stderr about files that were not changed or were " + "ignored due to --exclude=." ), ) @click.version_option(version=__version__) @@ -199,43 +267,51 @@ def main( fast: bool, pyi: bool, py36: bool, + skip_string_normalization: bool, quiet: bool, + verbose: bool, + include: str, + exclude: str, src: List[str], ) -> None: """The uncompromising code formatter.""" - sources: List[Path] = [] + write_back = WriteBack.from_configuration(check=check, diff=diff) + mode = FileMode.from_configuration( + py36=py36, pyi=pyi, skip_string_normalization=skip_string_normalization + ) + report = Report(check=check, quiet=quiet, verbose=verbose) + sources: Set[Path] = set() + try: + include_regex = re.compile(include) + except re.error: + err(f"Invalid regular expression for include given: {include!r}") + ctx.exit(2) + try: + exclude_regex = re.compile(exclude) + except re.error: + err(f"Invalid regular expression for exclude given: {exclude!r}") + ctx.exit(2) + root = find_project_root(src) for s in src: p = Path(s) if p.is_dir(): - sources.extend(gen_python_files_in_dir(p)) - elif p.is_file(): + sources.update( + gen_python_files_in_dir(p, root, include_regex, exclude_regex, report) + ) + elif p.is_file() or s == "-": # if a file was explicitly given, we don't care about its extension - sources.append(p) - elif s == "-": - sources.append(Path("-")) + sources.add(p) else: err(f"invalid path: {s}") - - if check and not diff: - write_back = WriteBack.NO - elif diff: - write_back = WriteBack.DIFF - else: - write_back = WriteBack.YES - mode = FileMode.AUTO_DETECT - if py36: - mode |= FileMode.PYTHON36 - if pyi: - mode |= FileMode.PYI - report = Report(check=check, quiet=quiet) if len(sources) == 0: - out("No paths given. Nothing to do 😴") + if verbose or not quiet: + out("No paths given. Nothing to do 😴") ctx.exit(0) return elif len(sources) == 1: reformat_one( - src=sources[0], + src=sources.pop(), line_length=line_length, fast=fast, write_back=write_back, @@ -260,9 +336,9 @@ def main( ) finally: shutdown(loop) - if not quiet: - out("All done! ✨ 🍰 ✨") - click.echo(str(report)) + if verbose or not quiet: + out("All done! ✨ 🍰 ✨") + click.echo(str(report)) ctx.exit(report.return_code) @@ -291,8 +367,8 @@ def reformat_one( cache: Cache = {} if write_back != WriteBack.DIFF: cache = read_cache(line_length, mode) - src = src.resolve() - if src in cache and cache[src] == get_cache_info(src): + res_src = src.resolve() + if res_src in cache and cache[res_src] == get_cache_info(res_src): changed = Changed.CACHED if changed is not Changed.CACHED and format_file_in_place( src, @@ -310,7 +386,7 @@ def reformat_one( async def schedule_formatting( - sources: List[Path], + sources: Set[Path], line_length: int, fast: bool, write_back: WriteBack, @@ -330,7 +406,7 @@ async def schedule_formatting( if write_back != WriteBack.DIFF: cache = read_cache(line_length, mode) sources, cached = filter_cached(cache, sources) - for src in cached: + for src in sorted(cached): report.done(src, Changed.CACHED) cancelled = [] formatted = [] @@ -393,8 +469,10 @@ def format_file_in_place( """ if src.suffix == ".pyi": mode |= FileMode.PYI - with tokenize.open(src) as src_buffer: - src_contents = src_buffer.read() + + then = datetime.utcfromtimestamp(src.stat().st_mtime) + with open(src, "rb") as buf: + src_contents, encoding, newline = decode_bytes(buf.read()) try: dst_contents = format_file_contents( src_contents, line_length=line_length, fast=fast, mode=mode @@ -403,16 +481,24 @@ def format_file_in_place( return False if write_back == write_back.YES: - with open(src, "w", encoding=src_buffer.encoding) as f: + with open(src, "w", encoding=encoding, newline=newline) as f: f.write(dst_contents) elif write_back == write_back.DIFF: - src_name = f"{src} (original)" - dst_name = f"{src} (formatted)" + now = datetime.utcnow() + 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: - sys.stdout.write(diff_contents) + f = io.TextIOWrapper( + sys.stdout.buffer, + encoding=encoding, + newline=newline, + write_through=True, + ) + f.write(diff_contents) + f.detach() finally: if lock: lock.release() @@ -431,7 +517,8 @@ def format_stdin_to_stdout( `line_length`, `fast`, `is_pyi`, and `force_py36` arguments are passed to :func:`format_file_contents`. """ - src = sys.stdin.read() + then = datetime.utcnow() + src, encoding, newline = decode_bytes(sys.stdin.buffer.read()) dst = src try: dst = format_file_contents(src, line_length=line_length, fast=fast, mode=mode) @@ -441,12 +528,17 @@ def format_stdin_to_stdout( return False finally: + f = io.TextIOWrapper( + sys.stdout.buffer, encoding=encoding, newline=newline, write_through=True + ) if write_back == WriteBack.YES: - sys.stdout.write(dst) + f.write(dst) elif write_back == WriteBack.DIFF: - src_name = " (original)" - dst_name = " (formatted)" - sys.stdout.write(diff(src, dst, src_name, dst_name)) + now = datetime.utcnow() + src_name = f"STDIN\t{then} +0000" + dst_name = f"STDOUT\t{now} +0000" + f.write(diff(src, dst, src_name, dst_name)) + f.detach() def format_file_contents( @@ -487,8 +579,11 @@ def format_str( future_imports = get_future_imports(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() @@ -504,6 +599,23 @@ def format_str( return dst_contents +def decode_bytes(src: bytes) -> Tuple[FileContent, Encoding, NewLine]: + """Return a tuple of (decoded_contents, encoding, newline). + + `newline` is either CRLF or LF but `decoded_contents` is decoded with + universal newlines (i.e. only contains LF). + """ + srcbuf = io.BytesIO(src) + encoding, lines = tokenize.detect_encoding(srcbuf.readline) + if not lines: + return "", encoding, "\n" + + newline = "\r\n" if b"\r\n" == lines[0][-2:] else "\n" + srcbuf.seek(0) + with io.TextIOWrapper(srcbuf, encoding) as tiow: + return tiow.read(), encoding, newline + + GRAMMARS = [ pygram.python_grammar_no_print_statement_no_exec_statement, pygram.python_grammar_no_print_statement, @@ -514,9 +626,8 @@ GRAMMARS = [ def lib2to3_parse(src_txt: str) -> Node: """Given a string with source, return the lib2to3 Node.""" grammar = pygram.python_grammar_no_print_statement - if src_txt[-1] != "\n": - nl = "\r\n" if "\r\n" in src_txt[:1024] else "\n" - src_txt += nl + if src_txt[-1:] != "\n": + src_txt += "\n" for grammar in GRAMMARS: drv = driver.Driver(grammar, pytree.convert) try: @@ -660,6 +771,7 @@ UNPACKING_PARENTS = { syms.dictsetmaker, syms.listmaker, syms.testlist_gexp, + syms.testlist_star_expr, } TEST_DESCENDANTS = { syms.test, @@ -982,6 +1094,13 @@ class Line: return False + def contains_multiline_strings(self) -> bool: + for leaf in self.leaves: + if is_multiline_string(leaf): + return True + + return False + def maybe_remove_trailing_comma(self, closing: Leaf) -> bool: """Remove trailing comma if there is one and it's safe.""" if not ( @@ -1286,6 +1405,7 @@ class LineGenerator(Visitor[Line]): """ is_pyi: bool = False + normalize_strings: bool = True current_line: Line = Factory(Line) remove_u_prefix: bool = False @@ -1354,7 +1474,7 @@ 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: @@ -2100,32 +2220,50 @@ def right_hand_split( result.append(leaf, preformatted=True) for comment_after in line.comments_after(leaf): result.append(comment_after, preformatted=True) - bracket_split_succeeded_or_raise(head, body, tail) assert opening_bracket and closing_bracket + body.should_explode = should_explode(body, opening_bracket) + bracket_split_succeeded_or_raise(head, body, tail) if ( + # the body shouldn't be exploded + not body.should_explode # the opening bracket is an optional paren - opening_bracket.type == token.LPAR + and opening_bracket.type == token.LPAR and not opening_bracket.value # the closing bracket is an optional paren and closing_bracket.type == token.RPAR and not closing_bracket.value - # 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) + # 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 + # there are no standalone comments in the body + and not body.contains_standalone_comments(0) + # and we can actually remove the parens + and can_omit_invisible_parens(body, line_length) ): omit = {id(closing_bracket), *omit} - if can_omit_invisible_parens(body, line_length): - try: - yield from right_hand_split(line, line_length, py36=py36, omit=omit) - return - except CannotSplit: - pass + try: + yield from right_hand_split(line, line_length, py36=py36, omit=omit) + return + + except CannotSplit: + if not ( + can_be_split(body) + or is_line_short_enough(body, line_length=line_length) + ): + raise CannotSplit( + "Splitting failed, body is still too long and can't be split." + ) + + elif head.contains_multiline_strings() or tail.contains_multiline_strings(): + raise CannotSplit( + "The current optional pair of parentheses is bound to fail to " + "satisfy the splitting algorithm becase the head or the tail " + "contains multiline strings which by definition never fit one " + "line." + ) 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 @@ -2736,33 +2874,64 @@ def get_future_imports(node: Node) -> Set[str]: return imports -PYTHON_EXTENSIONS = {".py", ".pyi"} -BLACKLISTED_DIRECTORIES = { - "build", - "buck-out", - "dist", - "_build", - ".git", - ".hg", - ".mypy_cache", - ".tox", - ".venv", -} - +def gen_python_files_in_dir( + path: Path, + root: Path, + include: Pattern[str], + exclude: Pattern[str], + report: "Report", +) -> Iterator[Path]: + """Generate all files under `path` whose paths are not excluded by the + `exclude` regex, but are included by the `include` regex. -def gen_python_files_in_dir(path: Path) -> Iterator[Path]: - """Generate all files under `path` which aren't under BLACKLISTED_DIRECTORIES - and have one of the PYTHON_EXTENSIONS. + `report` is where output about exclusions goes. """ + assert root.is_absolute(), f"INTERNAL ERROR: `root` must be absolute but is {root}" for child in path.iterdir(): + normalized_path = "/" + child.resolve().relative_to(root).as_posix() if child.is_dir(): - if child.name in BLACKLISTED_DIRECTORIES: - continue + normalized_path += "/" + exclude_match = exclude.search(normalized_path) + if exclude_match and exclude_match.group(0): + report.path_ignored(child, f"matches --exclude={exclude.pattern}") + continue + + if child.is_dir(): + yield from gen_python_files_in_dir(child, root, include, exclude, report) + + elif child.is_file(): + include_match = include.search(normalized_path) + if include_match: + yield child + + +def find_project_root(srcs: List[str]) -> Path: + """Return a directory containing .git, .hg, or pyproject.toml. + + That directory can be one of the directories passed in `srcs` or their + common parent. + + If no directory in the tree contains a marker that would specify it's the + project root, the root of the file system is returned. + """ + if not srcs: + return Path("/").resolve() + + common_base = min(Path(src).resolve() for src in srcs) + if common_base.is_dir(): + # Append a fake file so `parents` below returns `common_base_dir`, too. + common_base /= "fake-file" + for directory in common_base.parents: + if (directory / ".git").is_dir(): + return directory - yield from gen_python_files_in_dir(child) + if (directory / ".hg").is_dir(): + return directory - elif child.is_file() and child.suffix in PYTHON_EXTENSIONS: - yield child + if (directory / "pyproject.toml").is_file(): + return directory + + return directory @dataclass @@ -2771,6 +2940,7 @@ class Report: check: bool = False quiet: bool = False + verbose: bool = False change_count: int = 0 same_count: int = 0 failure_count: int = 0 @@ -2779,11 +2949,11 @@ class Report: """Increment the counter for successful reformatting. Write out a message.""" if changed is Changed.YES: reformatted = "would reformat" if self.check else "reformatted" - if not self.quiet: + if self.verbose or not self.quiet: out(f"{reformatted} {src}") self.change_count += 1 else: - if not self.quiet: + if self.verbose: if changed is Changed.NO: msg = f"{src} already well formatted, good job." else: @@ -2796,6 +2966,10 @@ class Report: err(f"error: cannot format {src}: {message}") self.failure_count += 1 + def path_ignored(self, path: Path, message: str) -> None: + if self.verbose: + out(f"{path} ignored: {message}", bold=False) + @property def return_code(self) -> int: """Return the exit code that the app should use. @@ -3034,6 +3208,42 @@ def is_line_short_enough(line: Line, *, line_length: int, line_str: str = "") -> ) +def can_be_split(line: Line) -> bool: + """Return False if the line cannot be split *for sure*. + + This is not an exhaustive search but a cheap heuristic that we can use to + avoid some unfortunate formattings (mostly around wrapping unsplittable code + in unnecessary parentheses). + """ + leaves = line.leaves + if len(leaves) < 2: + return False + + if leaves[0].type == token.STRING and leaves[1].type == token.DOT: + call_count = 0 + dot_count = 0 + next = leaves[-1] + for leaf in leaves[-2::-1]: + if leaf.type in OPENING_BRACKETS: + if next.type not in CLOSING_BRACKETS: + return False + + call_count += 1 + elif leaf.type == token.DOT: + dot_count += 1 + elif leaf.type == token.NAME: + if not (next.type == token.DOT or next.type in OPENING_BRACKETS): + return False + + elif leaf.type not in CLOSING_BRACKETS: + return False + + if dot_count > 1 and call_count > 1: + return False + + return True + + def can_omit_invisible_parens(line: Line, line_length: int) -> bool: """Does `line` have a shape safe to reformat without optional parens around it? @@ -3156,26 +3366,24 @@ def get_cache_info(path: Path) -> CacheInfo: return stat.st_mtime, stat.st_size -def filter_cached( - cache: Cache, sources: Iterable[Path] -) -> Tuple[List[Path], List[Path]]: - """Split a list of paths into two. +def filter_cached(cache: Cache, sources: Iterable[Path]) -> Tuple[Set[Path], Set[Path]]: + """Split an iterable of paths in `sources` into two sets. - The first list contains paths of files that modified on disk or are not in the - cache. The other list contains paths to non-modified files. + The first contains paths of files that modified on disk or are not in the + cache. The other contains paths to non-modified files. """ - todo, done = [], [] + todo, done = set(), set() for src in sources: src = src.resolve() if cache.get(src) != get_cache_info(src): - todo.append(src) + todo.add(src) else: - done.append(src) + done.add(src) return todo, done def write_cache( - cache: Cache, sources: List[Path], line_length: int, mode: FileMode + cache: Cache, sources: Iterable[Path], line_length: int, mode: FileMode ) -> None: """Update the cache file.""" cache_file = get_cache_file(line_length, mode)