X-Git-Url: https://git.madduck.net/etc/vim.git/blobdiff_plain/e911c79809c4fd9b0773dea5b6a0e710b59614cf..c98a6f134f4b2b831e7e659de0c217c118270a61:/black.py?ds=sidebyside diff --git a/black.py b/black.py index 8a4eb05..f03c4f1 100644 --- a/black.py +++ b/black.py @@ -1,4 +1,5 @@ #!/usr/bin/env python3 + import asyncio from asyncio.base_events import BaseEventLoop from concurrent.futures import Executor, ProcessPoolExecutor @@ -21,7 +22,7 @@ from blib2to3 import pygram, pytree from blib2to3.pgen2 import driver, token from blib2to3.pgen2.parse import ParseError -__version__ = "18.3a2" +__version__ = "18.3a3" DEFAULT_LINE_LENGTH = 88 # types syms = pygram.python_symbols @@ -43,7 +44,7 @@ class NothingChanged(UserWarning): class CannotSplit(Exception): """A readable split that fits the allotted line length is impossible. - Raised by `left_hand_split()` and `right_hand_split()`. + Raised by `left_hand_split()`, `right_hand_split()`, and `delimiter_split()`. """ @@ -61,7 +62,7 @@ class CannotSplit(Exception): is_flag=True, help=( "Don't write back the files, just return the status. Return code 0 " - "means nothing changed. Return code 1 means some files were " + "means nothing would change. Return code 1 means some files would be " "reformatted. Return code 123 means there was an internal error." ), ) @@ -99,10 +100,12 @@ def main( ctx.exit(0) elif len(sources) == 1: p = sources[0] - report = Report() + report = Report(check=check) try: if not p.is_file() and str(p) == '-': - changed = format_stdin_to_stdout(line_length=line_length, fast=fast) + changed = format_stdin_to_stdout( + line_length=line_length, fast=fast, write_back=not check + ) else: changed = format_file_in_place( p, line_length=line_length, fast=fast, write_back=not check @@ -178,7 +181,9 @@ def format_file_in_place( return True -def format_stdin_to_stdout(line_length: int, fast: bool) -> bool: +def format_stdin_to_stdout( + line_length: int, fast: bool, write_back: bool = False +) -> bool: """Format file on stdin and pipe output to stdout. Return True if changed.""" contents = sys.stdin.read() try: @@ -189,7 +194,8 @@ def format_stdin_to_stdout(line_length: int, fast: bool) -> bool: return False finally: - sys.stdout.write(contents) + if write_back: + sys.stdout.write(contents) def format_file_contents( @@ -213,7 +219,6 @@ def format_str(src_contents: str, line_length: int) -> FileContent: """Reformats a string and returns new contents.""" src_node = lib2to3_parse(src_contents) dst_contents = "" - comments: List[Line] = [] lines = LineGenerator() elt = EmptyLineTracker() py36 = is_python36(src_node) @@ -225,41 +230,41 @@ def format_str(src_contents: str, line_length: int) -> FileContent: before, after = elt.maybe_empty_lines(current_line) for _ in range(before): dst_contents += str(empty_line) - if not current_line.is_comment: - for comment in comments: - dst_contents += str(comment) - comments = [] - for line in split_line(current_line, line_length=line_length, py36=py36): - dst_contents += str(line) - else: - comments.append(current_line) - if comments: - if elt.previous_defs: - # Separate postscriptum comments from the last module-level def. - dst_contents += str(empty_line) - dst_contents += str(empty_line) - for comment in comments: - dst_contents += str(comment) + for line in split_line(current_line, line_length=line_length, py36=py36): + dst_contents += str(line) return dst_contents +GRAMMARS = [ + pygram.python_grammar_no_print_statement_no_exec_statement, + pygram.python_grammar_no_print_statement, + pygram.python_grammar_no_exec_statement, + pygram.python_grammar, +] + + def lib2to3_parse(src_txt: str) -> Node: """Given a string with source, return the lib2to3 Node.""" grammar = pygram.python_grammar_no_print_statement - drv = driver.Driver(grammar, pytree.convert) if src_txt[-1] != '\n': nl = '\r\n' if '\r\n' in src_txt[:1024] else '\n' src_txt += nl - try: - result = drv.parse_string(src_txt, True) - except ParseError as pe: - lineno, column = pe.context[1] - lines = src_txt.splitlines() + for grammar in GRAMMARS: + drv = driver.Driver(grammar, pytree.convert) try: - faulty_line = lines[lineno - 1] - except IndexError: - faulty_line = "" - raise ValueError(f"Cannot parse: {lineno}:{column}: {faulty_line}") from None + result = drv.parse_string(src_txt, True) + break + + except ParseError as pe: + lineno, column = pe.context[1] + lines = src_txt.splitlines() + try: + faulty_line = lines[lineno - 1] + except IndexError: + faulty_line = "" + exc = ValueError(f"Cannot parse: {lineno}:{column}: {faulty_line}") + else: + raise exc from None if isinstance(result, Leaf): result = Node(syms.file_input, [result]) @@ -348,6 +353,7 @@ MATH_OPERATORS = { token.AMPER, token.PERCENT, token.CIRCUMFLEX, + token.TILDE, token.LEFTSHIFT, token.RIGHTSHIFT, token.DOUBLESTAR, @@ -413,7 +419,7 @@ class BracketTracker: """Returns True if there is an yet unmatched open bracket on the line.""" return bool(self.bracket_match) - def max_priority(self, exclude: Iterable[LeafID] =()) -> int: + def max_priority(self, exclude: Iterable[LeafID] = ()) -> int: """Returns the highest priority of a delimiter found on the line. Values are consistent with what `is_delimiter()` returns. @@ -517,10 +523,16 @@ class Line: ): return False - if closing.type == token.RSQB or closing.type == token.RBRACE: + if closing.type == token.RBRACE: self.leaves.pop() return True + if closing.type == token.RSQB: + comma = self.leaves[-1] + if comma.parent and comma.parent.type == syms.listmaker: + self.leaves.pop() + return True + # For parens let's check if it's safe to remove the comma. If the # trailing one is the only one, we might mistakenly change a tuple # into a different type by removing the comma. @@ -657,10 +669,6 @@ class EmptyLineTracker: (two on module-level), as well as providing an extra empty line after flow control keywords to make them more prominent. """ - if current_line.is_comment: - # Don't count standalone comments towards previous empty lines. - return 0, 0 - before, after = self._maybe_empty_lines(current_line) before -= self.previous_after self.previous_after = after @@ -668,10 +676,14 @@ class EmptyLineTracker: return before, after def _maybe_empty_lines(self, current_line: Line) -> Tuple[int, int]: + max_allowed = 1 + if current_line.is_comment and current_line.depth == 0: + max_allowed = 2 if current_line.leaves: # Consume the first leaf's extra newlines. first_leaf = current_line.leaves[0] - before = int('\n' in first_leaf.prefix) + before = first_leaf.prefix.count('\n') + before = min(before, max(before, max_allowed)) first_leaf.prefix = '' else: before = 0 @@ -725,7 +737,6 @@ class LineGenerator(Visitor[Line]): in ways that will no longer stringify to valid Python code on the tree. """ current_line: Line = Factory(Line) - standalone_comments: List[Leaf] = Factory(list) def line(self, indent: int = 0) -> Iterator[Line]: """Generate a line. @@ -756,33 +767,22 @@ class LineGenerator(Visitor[Line]): yield from self.line() else: - # regular standalone comment, to be processed later (see - # docstring in `generate_comments()` - self.standalone_comments.append(comment) - normalize_prefix(node, inside_brackets=any_open_brackets) - if node.type not in WHITESPACE: - for comment in self.standalone_comments: + # regular standalone comment yield from self.line() self.current_line.append(comment) yield from self.line() - self.standalone_comments = [] + normalize_prefix(node, inside_brackets=any_open_brackets) + if node.type not in WHITESPACE: self.current_line.append(node) yield from super().visit_default(node) - def visit_suite(self, node: Node) -> Iterator[Line]: - """Body of a statement after a colon.""" - children = iter(node.children) - # Process newline before indenting. It might contain an inline - # comment that should go right after the colon. - newline = next(children) - yield from self.visit(newline) + def visit_INDENT(self, node: Node) -> Iterator[Line]: yield from self.line(+1) + yield from self.visit_default(node) - for child in children: - yield from self.visit(child) - + def visit_DEDENT(self, node: Node) -> Iterator[Line]: yield from self.line(-1) def visit_stmt(self, node: Node, keywords: Set[str]) -> Iterator[Line]: @@ -872,7 +872,7 @@ def whitespace(leaf: Leaf) -> 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 != syms.subscript: + if t == token.COLON and p.type not in {syms.subscript, syms.subscriptlist}: return NO prev = leaf.prev_sibling @@ -885,30 +885,49 @@ def whitespace(leaf: Leaf) -> str: # noqa C901 return SPACE if prevp.type == token.COMMA else NO if prevp.type == token.EQUAL: - if prevp.parent and prevp.parent.type in { - syms.typedargslist, - syms.varargslist, - syms.parameters, - syms.arglist, - syms.argument, - }: - return NO + if prevp.parent: + if prevp.parent.type in { + syms.arglist, syms.argument, syms.parameters, syms.varargslist + }: + return NO + + elif prevp.parent.type == syms.typedargslist: + # A bit hacky: if the equal sign has whitespace, it means we + # previously found it's a typed argument. So, we're using + # that, too. + return prevp.prefix elif prevp.type == token.DOUBLESTAR: if prevp.parent and prevp.parent.type in { - syms.typedargslist, - syms.varargslist, - syms.parameters, syms.arglist, + syms.argument, syms.dictsetmaker, + syms.parameters, + syms.typedargslist, + syms.varargslist, }: return NO elif prevp.type == token.COLON: - if prevp.parent and prevp.parent.type == syms.subscript: + if prevp.parent and prevp.parent.type in {syms.subscript, syms.sliceop}: return NO - elif prevp.parent and prevp.parent.type in {syms.factor, syms.star_expr}: + elif ( + prevp.parent + and prevp.parent.type in {syms.factor, syms.star_expr} + and prevp.type in MATH_OPERATORS + ): + return NO + + elif ( + prevp.type == token.RIGHTSHIFT + and prevp.parent + and prevp.parent.type == syms.shift_expr + and prevp.prev_sibling + and prevp.prev_sibling.type == token.NAME + and prevp.prev_sibling.value == 'print' # type: ignore + ): + # Python 2 print chevron return NO elif prev.type in OPENING_BRACKETS: @@ -922,7 +941,7 @@ def whitespace(leaf: Leaf) -> str: # noqa C901 if not prev or prev.type != token.COMMA: return NO - if p.type == syms.varargslist: + elif p.type == syms.varargslist: # lambdas if t == token.RPAR: return NO @@ -1133,30 +1152,40 @@ def generate_comments(leaf: Leaf) -> Iterator[Leaf]: Inline comments are emitted as regular token.COMMENT leaves. Standalone are emitted with a fake STANDALONE_COMMENT token identifier. """ - if not leaf.prefix: + p = leaf.prefix + if not p: return - if '#' not in leaf.prefix: + if '#' not in p: return - before_comment, content = leaf.prefix.split('#', 1) - content = content.rstrip() - if content and (content[0] not in {' ', '!', '#'}): - content = ' ' + content - is_standalone_comment = ( - '\n' in before_comment or '\n' in content or leaf.type == token.ENDMARKER - ) - if not is_standalone_comment: - # simple trailing comment - yield Leaf(token.COMMENT, value='#' + content) - return - - for line in ('#' + content).split('\n'): + nlines = 0 + for index, line in enumerate(p.split('\n')): line = line.lstrip() + if not line: + nlines += 1 if not line.startswith('#'): continue - yield Leaf(STANDALONE_COMMENT, line) + if index == 0 and leaf.type != token.ENDMARKER: + comment_type = token.COMMENT # simple trailing comment + else: + comment_type = STANDALONE_COMMENT + yield Leaf(comment_type, make_comment(line), prefix='\n' * nlines) + + nlines = 0 + + +def make_comment(content: str) -> str: + content = content.rstrip() + if not content: + return '#' + + if content[0] == '#': + content = content[1:] + if content and content[0] not in ' !:#': + content = ' ' + content + return '#' + content def split_line( @@ -1379,9 +1408,11 @@ def normalize_prefix(leaf: Leaf, *, inside_brackets: bool) -> None: you'll lose your voting rights. """ if not inside_brackets: - spl = leaf.prefix.split('#', 1) + spl = leaf.prefix.split('#') if '\\' not in spl[0]: - nl_count = spl[0].count('\n') + nl_count = spl[-1].count('\n') + if len(spl) > 1: + nl_count -= 1 leaf.prefix = '\n' * nl_count return @@ -1434,6 +1465,7 @@ def gen_python_files_in_dir(path: Path) -> Iterator[Path]: @dataclass class Report: """Provides a reformatting counter.""" + check: bool = False change_count: int = 0 same_count: int = 0 failure_count: int = 0 @@ -1441,7 +1473,8 @@ class Report: def done(self, src: Path, changed: bool) -> None: """Increment the counter for successful reformatting. Write out a message.""" if changed: - out(f'reformatted {src}') + reformatted = 'would reformat' if self.check else 'reformatted' + out(f'{reformatted} {src}') self.change_count += 1 else: out(f'{src} already well formatted, good job.', bold=False) @@ -1460,7 +1493,7 @@ class Report: if self.failure_count: return 123 - elif self.change_count: + elif self.change_count and self.check: return 1 return 0 @@ -1470,21 +1503,27 @@ class Report: Use `click.unstyle` to remove colors. """ + if self.check: + reformatted = "would be reformatted" + unchanged = "would be left unchanged" + failed = "would fail to reformat" + else: + reformatted = "reformatted" + unchanged = "left unchanged" + failed = "failed to reformat" report = [] if self.change_count: s = 's' if self.change_count > 1 else '' report.append( - click.style(f'{self.change_count} file{s} reformatted', bold=True) + click.style(f'{self.change_count} file{s} {reformatted}', bold=True) ) if self.same_count: s = 's' if self.same_count > 1 else '' - report.append(f'{self.same_count} file{s} left unchanged') + report.append(f'{self.same_count} file{s} {unchanged}') if self.failure_count: s = 's' if self.failure_count > 1 else '' report.append( - click.style( - f'{self.failure_count} file{s} failed to reformat', fg='red' - ) + click.style(f'{self.failure_count} file{s} {failed}', fg='red') ) return ', '.join(report) + '.' @@ -1526,7 +1565,12 @@ def assert_equivalent(src: str, dst: str) -> None: try: src_ast = ast.parse(src) except Exception as exc: - raise AssertionError(f"cannot parse source: {exc}") from None + major, minor = sys.version_info[:2] + raise AssertionError( + f"cannot use --safe with this file; failed to parse source file " + f"with Python {major}.{minor}'s builtin AST. Re-run with --fast " + f"or stop using deprecated Python 2 syntax. AST error message: {exc}" + ) try: dst_ast = ast.parse(dst)