#!/usr/bin/env python3
+
import asyncio
from asyncio.base_events import BaseEventLoop
from concurrent.futures import Executor, ProcessPoolExecutor
Dict, Generic, Iterable, Iterator, List, Optional, Set, Tuple, TypeVar, Union
)
-from attr import attrib, dataclass, Factory
+from attr import dataclass, Factory
import click
# lib2to3 fork
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
@click.argument(
'src',
nargs=-1,
- type=click.Path(exists=True, file_okay=True, dir_okay=True, readable=True),
+ type=click.Path(
+ exists=True, file_okay=True, dir_okay=True, readable=True, allow_dash=True
+ ),
)
@click.pass_context
def main(
elif p.is_file():
# if a file was explicitly given, we don't care about its extension
sources.append(p)
+ elif s == '-':
+ sources.append(Path('-'))
else:
err(f'invalid path: {s}')
if len(sources) == 0:
p = sources[0]
report = Report()
try:
- changed = format_file_in_place(
- p, line_length=line_length, fast=fast, write_back=not check
- )
+ if not p.is_file() and str(p) == '-':
+ 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
+ )
report.done(p, changed)
except Exception as exc:
report.failed(p, str(exc))
src: Path, line_length: int, fast: bool, write_back: bool = False
) -> bool:
"""Format the file and rewrite if changed. Return True if changed."""
+ with tokenize.open(src) as src_buffer:
+ src_contents = src_buffer.read()
try:
- contents, encoding = format_file(src, line_length=line_length, fast=fast)
+ contents = format_file_contents(
+ src_contents, line_length=line_length, fast=fast
+ )
except NothingChanged:
return False
if write_back:
- with open(src, "w", encoding=encoding) as f:
+ with open(src, "w", encoding=src_buffer.encoding) as f:
f.write(contents)
return True
-def format_file(
- src: Path, line_length: int, fast: bool
-) -> Tuple[FileContent, Encoding]:
+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:
+ contents = format_file_contents(contents, line_length=line_length, fast=fast)
+ return True
+
+ except NothingChanged:
+ return False
+
+ finally:
+ if write_back:
+ sys.stdout.write(contents)
+
+
+def format_file_contents(
+ src_contents: str, line_length: int, fast: bool
+) -> FileContent:
"""Reformats a file and returns its contents and encoding."""
- with tokenize.open(src) as src_buffer:
- src_contents = src_buffer.read()
if src_contents.strip() == '':
- raise NothingChanged(src)
+ raise NothingChanged
dst_contents = format_str(src_contents, line_length=line_length)
if src_contents == dst_contents:
- raise NothingChanged(src)
+ raise NothingChanged
if not fast:
assert_equivalent(src_contents, dst_contents)
assert_stable(src_contents, dst_contents, line_length=line_length)
- return dst_contents, src_buffer.encoding
+ return dst_contents
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)
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)
- 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
@dataclass
class DebugVisitor(Visitor[T]):
- tree_depth: int = attrib(default=0)
+ tree_depth: int = 0
def visit_default(self, node: LN) -> Iterator[T]:
indent = ' ' * (2 * self.tree_depth)
token.AMPER,
token.PERCENT,
token.CIRCUMFLEX,
+ token.TILDE,
token.LEFTSHIFT,
token.RIGHTSHIFT,
token.DOUBLESTAR,
@dataclass
class BracketTracker:
- depth: int = attrib(default=0)
- bracket_match: Dict[Tuple[Depth, NodeType], Leaf] = attrib(default=Factory(dict))
- delimiters: Dict[LeafID, Priority] = attrib(default=Factory(dict))
- previous: Optional[Leaf] = attrib(default=None)
+ depth: int = 0
+ bracket_match: Dict[Tuple[Depth, NodeType], Leaf] = Factory(dict)
+ delimiters: Dict[LeafID, Priority] = Factory(dict)
+ previous: Optional[Leaf] = None
def mark(self, leaf: Leaf) -> None:
if leaf.type == token.COMMENT:
if leaf.type == token.STRING and self.previous.type == token.STRING:
self.delimiters[id(self.previous)] = STRING_PRIORITY
elif (
- leaf.type == token.NAME and
- leaf.value == 'for' and
- leaf.parent and
- leaf.parent.type in {syms.comp_for, syms.old_comp_for}
+ leaf.type == token.NAME
+ and leaf.value == 'for'
+ and leaf.parent
+ and leaf.parent.type in {syms.comp_for, syms.old_comp_for}
):
self.delimiters[id(self.previous)] = COMPREHENSION_PRIORITY
elif (
- leaf.type == token.NAME and
- leaf.value == 'if' and
- leaf.parent and
- leaf.parent.type in {syms.comp_if, syms.old_comp_if}
+ leaf.type == token.NAME
+ and leaf.value == 'if'
+ and leaf.parent
+ and leaf.parent.type in {syms.comp_if, syms.old_comp_if}
):
self.delimiters[id(self.previous)] = COMPREHENSION_PRIORITY
+ elif (
+ leaf.type == token.NAME
+ and leaf.value in LOGIC_OPERATORS
+ and leaf.parent
+ ):
+ self.delimiters[id(self.previous)] = LOGIC_PRIORITY
if leaf.type in OPENING_BRACKETS:
self.bracket_match[self.depth, BRACKET[leaf.type]] = leaf
self.depth += 1
@dataclass
class Line:
- depth: int = attrib(default=0)
- leaves: List[Leaf] = attrib(default=Factory(list))
- comments: Dict[LeafID, Leaf] = attrib(default=Factory(dict))
- bracket_tracker: BracketTracker = attrib(default=Factory(BracketTracker))
- inside_brackets: bool = attrib(default=False)
- has_for: bool = attrib(default=False)
- _for_loop_variable: bool = attrib(default=False, init=False)
+ depth: int = 0
+ leaves: List[Leaf] = Factory(list)
+ comments: Dict[LeafID, Leaf] = Factory(dict)
+ bracket_tracker: BracketTracker = Factory(BracketTracker)
+ inside_brackets: bool = False
+ has_for: bool = False
+ _for_loop_variable: bool = False
def append(self, leaf: Leaf, preformatted: bool = False) -> None:
has_value = leaf.value.strip()
@property
def is_class(self) -> bool:
return (
- bool(self) and
- self.leaves[0].type == token.NAME and
- self.leaves[0].value == 'class'
+ bool(self)
+ and self.leaves[0].type == token.NAME
+ and self.leaves[0].value == 'class'
)
@property
except IndexError:
second_leaf = None
return (
- (first_leaf.type == token.NAME and first_leaf.value == 'def') or
- (
- first_leaf.type == token.NAME and
- first_leaf.value == 'async' and
- second_leaf is not None and
- second_leaf.type == token.NAME and
- second_leaf.value == 'def'
+ (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
def is_flow_control(self) -> bool:
return (
- bool(self) and
- self.leaves[0].type == token.NAME and
- self.leaves[0].value in FLOW_CONTROL
+ bool(self)
+ and self.leaves[0].type == token.NAME
+ and self.leaves[0].value in FLOW_CONTROL
)
@property
def is_yield(self) -> bool:
return (
- bool(self) and
- self.leaves[0].type == token.NAME and
- self.leaves[0].value == 'yield'
+ bool(self)
+ and self.leaves[0].type == token.NAME
+ and self.leaves[0].value == 'yield'
)
def maybe_remove_trailing_comma(self, closing: Leaf) -> bool:
if not (
- self.leaves and
- self.leaves[-1].type == token.COMMA and
- closing.type in CLOSING_BRACKETS
+ self.leaves
+ and self.leaves[-1].type == token.COMMA
+ and closing.type in CLOSING_BRACKETS
):
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.
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:
+ commas += 1
+ break
+
if commas > 1:
self.leaves.pop()
return True
appended will appear "too long" when splitting.
"""
if not (
- comment.type == STANDALONE_COMMENT and
- self.bracket_tracker.any_open_brackets()
+ comment.type == STANDALONE_COMMENT
+ and self.bracket_tracker.any_open_brackets()
):
return False
"""Provides a stateful method that returns the number of potential extra
empty lines needed before and after the currently processed line.
- Note: this tracker works on lines that haven't been split yet.
+ Note: this tracker works on lines that haven't been split yet. It assumes
+ the prefix of the first leaf consists of optional newlines. Those newlines
+ are consumed by `maybe_empty_lines()` and included in the computation.
"""
- previous_line: Optional[Line] = attrib(default=None)
- previous_after: int = attrib(default=0)
- previous_defs: List[int] = attrib(default=Factory(list))
+ previous_line: Optional[Line] = None
+ previous_after: int = 0
+ previous_defs: List[int] = Factory(list)
def maybe_empty_lines(self, current_line: Line) -> Tuple[int, int]:
"""Returns the number of extra empty lines before and after the `current_line`.
control keywords to make them more prominent.
"""
before, after = self._maybe_empty_lines(current_line)
+ before -= self.previous_after
self.previous_after = after
self.previous_line = current_line
return before, after
def _maybe_empty_lines(self, current_line: Line) -> Tuple[int, int]:
- before = 0
+ 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 = first_leaf.prefix.count('\n')
+ before = min(before, max(before, max_allowed))
+ first_leaf.prefix = ''
+ else:
+ before = 0
depth = current_line.depth
while self.previous_defs and self.previous_defs[-1] >= depth:
self.previous_defs.pop()
- before = (1 if depth else 2) - self.previous_after
+ before = 1 if depth else 2
is_decorator = current_line.is_decorator
if is_decorator or current_line.is_def or current_line.is_class:
if not is_decorator:
newlines = 2
if current_line.depth:
newlines -= 1
- newlines -= self.previous_after
return newlines, 0
if current_line.is_flow_control:
return before, 1
if (
- self.previous_line and
- self.previous_line.is_import and
- not current_line.is_import and
- depth == self.previous_line.depth
+ self.previous_line
+ and self.previous_line.is_import
+ and not current_line.is_import
+ and depth == self.previous_line.depth
):
return (before or 1), 0
if (
- self.previous_line and
- self.previous_line.is_yield and
- (not current_line.is_yield or depth != self.previous_line.depth)
+ self.previous_line
+ and self.previous_line.is_yield
+ and (not current_line.is_yield or depth != self.previous_line.depth)
):
return (before or 1), 0
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.
"""
- current_line: Line = attrib(default=Factory(Line))
- standalone_comments: List[Leaf] = attrib(default=Factory(list))
+ current_line: Line = Factory(Line)
def line(self, indent: int = 0) -> Iterator[Line]:
"""Generate a line.
def visit_default(self, node: LN) -> Iterator[Line]:
if isinstance(node, Leaf):
+ any_open_brackets = self.current_line.bracket_tracker.any_open_brackets()
for comment in generate_comments(node):
- if self.current_line.bracket_tracker.any_open_brackets():
+ if any_open_brackets:
# any comment within brackets is subject to splitting
self.current_line.append(comment)
elif comment.type == token.COMMENT:
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)
- 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]:
for child in children:
yield from self.visit(child)
- if child.type == token.NAME and child.value == 'async': # type: ignore
+ if child.type == token.ASYNC:
break
internal_stmt = next(children)
OPENING_BRACKETS = set(BRACKET.keys())
CLOSING_BRACKETS = set(BRACKET.values())
BRACKETS = OPENING_BRACKETS | CLOSING_BRACKETS
-ALWAYS_NO_SPACE = CLOSING_BRACKETS | {token.COMMA, token.COLON, STANDALONE_COMMENT}
+ALWAYS_NO_SPACE = CLOSING_BRACKETS | {token.COMMA, STANDALONE_COMMENT}
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 not in {syms.subscript, syms.subscriptlist}:
+ return NO
+
prev = leaf.prev_sibling
if not prev:
prevp = preceding_leaf(p)
if not prevp or prevp.type in OPENING_BRACKETS:
return NO
+ if t == token.COLON:
+ 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,
+ syms.parameters,
+ syms.typedargslist,
+ syms.varargslist,
}:
return NO
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 prev.type in OPENING_BRACKETS:
return NO
- elif prev.type == token.COLON:
+ else:
return NO
elif p.type == syms.atom:
return NO
elif (
- p.type == syms.listmaker or
- p.type == syms.testlist_gexp or
- p.type == syms.subscriptlist
+ p.type == syms.listmaker
+ or p.type == syms.testlist_gexp
+ or p.type == syms.subscriptlist
):
# list interior, including unpacking
if not prev:
if leaf.type == token.COMMA:
return COMMA_PRIORITY
- if leaf.type == token.NAME and leaf.value in LOGIC_OPERATORS:
- return LOGIC_PRIORITY
-
if leaf.type in COMPARATORS:
return COMPARATOR_PRIORITY
if (
- leaf.type in MATH_OPERATORS and
- leaf.parent and
- leaf.parent.type not in {syms.factor, syms.star_expr}
+ leaf.type in MATH_OPERATORS
+ and leaf.parent
+ and leaf.parent.type not in {syms.factor, syms.star_expr}
):
return MATH_PRIORITY
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.DEDENT
- )
- 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(
matching_bracket = None
for leaf in line.leaves:
if (
- current_leaves is body_leaves and
- leaf.type in CLOSING_BRACKETS and
- leaf.opening_bracket is matching_bracket
+ current_leaves is body_leaves
+ and leaf.type in CLOSING_BRACKETS
+ and leaf.opening_bracket is matching_bracket
):
current_leaves = tail_leaves if body_leaves else head_leaves
current_leaves.append(leaf)
current_leaves = body_leaves
# Since body is a new indent level, remove spurious leading whitespace.
if body_leaves:
- normalize_prefix(body_leaves[0])
+ normalize_prefix(body_leaves[0], inside_brackets=True)
# Build the new lines.
for result, leaves in (
(head, head_leaves), (body, body_leaves), (tail, tail_leaves)
head_leaves.reverse()
# Since body is a new indent level, remove spurious leading whitespace.
if body_leaves:
- normalize_prefix(body_leaves[0])
+ normalize_prefix(body_leaves[0], inside_brackets=True)
# Build the new lines.
for result, leaves in (
(head, head_leaves), (body, body_leaves), (tail, tail_leaves)
current_line.append(comment_after, preformatted=True)
lowest_depth = min(lowest_depth, leaf.bracket_depth)
if (
- leaf.bracket_depth == lowest_depth and
- leaf.type == token.STAR or
- leaf.type == token.DOUBLESTAR
+ leaf.bracket_depth == lowest_depth
+ and leaf.type == token.STAR
+ or leaf.type == token.DOUBLESTAR
):
trailing_comma_safe = trailing_comma_safe and py36
leaf_priority = delimiters.get(id(leaf))
if leaf_priority == delimiter_priority:
- normalize_prefix(current_line.leaves[0])
+ normalize_prefix(current_line.leaves[0], inside_brackets=True)
yield current_line
current_line = Line(depth=line.depth, inside_brackets=line.inside_brackets)
if current_line:
if (
- delimiter_priority == COMMA_PRIORITY and
- current_line.leaves[-1].type != token.COMMA and
- trailing_comma_safe
+ delimiter_priority == COMMA_PRIORITY
+ and current_line.leaves[-1].type != token.COMMA
+ and trailing_comma_safe
):
current_line.append(Leaf(token.COMMA, ','))
- normalize_prefix(current_line.leaves[0])
+ normalize_prefix(current_line.leaves[0], inside_brackets=True)
yield current_line
t = leaf.type
v = leaf.value
return bool(
- t == token.NAME and
- (
- (v == 'import' and p and p.type == syms.import_name) or
- (v == 'from' and p and p.type == syms.import_from)
+ t == token.NAME
+ and (
+ (v == 'import' and p and p.type == syms.import_name)
+ or (v == 'from' and p and p.type == syms.import_from)
)
)
-def normalize_prefix(leaf: Leaf) -> None:
- """Leave existing extra newlines for imports. Remove everything else."""
- if is_import(leaf):
- spl = leaf.prefix.split('#', 1)
- nl_count = spl[0].count('\n')
- if len(spl) > 1:
- # Skip one newline since it was for a standalone comment.
- nl_count -= 1
- leaf.prefix = '\n' * nl_count
- return
+def normalize_prefix(leaf: Leaf, *, inside_brackets: bool) -> None:
+ """Leave existing extra newlines if not `inside_brackets`.
+
+ Remove everything else. Note: don't use backslashes for formatting or
+ you'll lose your voting rights.
+ """
+ if not inside_brackets:
+ spl = leaf.prefix.split('#')
+ if '\\' not in spl[0]:
+ nl_count = spl[-1].count('\n')
+ if len(spl) > 1:
+ nl_count -= 1
+ leaf.prefix = '\n' * nl_count
+ return
leaf.prefix = ''
return True
elif (
- n.type == syms.typedargslist and
- n.children and
- n.children[-1].type == token.COMMA
+ n.type == syms.typedargslist
+ and n.children
+ and n.children[-1].type == token.COMMA
):
for ch in n.children:
if ch.type == token.STAR or ch.type == token.DOUBLESTAR:
@dataclass
class Report:
"""Provides a reformatting counter."""
- change_count: int = attrib(default=0)
- same_count: int = attrib(default=0)
- failure_count: int = attrib(default=0)
+ change_count: int = 0
+ same_count: int = 0
+ failure_count: int = 0
def done(self, src: Path, changed: bool) -> None:
"""Increment the counter for successful reformatting. Write out a message."""
raise AssertionError(
f"INTERNAL ERROR: Black produced invalid code: {exc}. "
f"Please report a bug on https://github.com/ambv/black/issues. "
- f"This invalid output might be helpful: {log}",
+ f"This invalid output might be helpful: {log}"
) from None
src_ast_str = '\n'.join(_v(src_ast))
f"INTERNAL ERROR: Black produced code that is not equivalent to "
f"the source. "
f"Please report a bug on https://github.com/ambv/black/issues. "
- f"This diff might be helpful: {log}",
+ f"This diff might be helpful: {log}"
) from None
f"INTERNAL ERROR: Black produced different code on the second pass "
f"of the formatter. "
f"Please report a bug on https://github.com/ambv/black/issues. "
- f"This diff might be helpful: {log}",
+ f"This diff might be helpful: {log}"
) from None
projects / etc / vim.git / blobdiff