Ignore symbolic links pointing outside of the root directory (#339)

[etc/vim.git] / black.py

diff --git a/black.py b/black.py
index 35af598e59032c62397bff766d06b006df41174c..e3b3882d993671bd8998e832946ef9b9df552c21 100644 (file)
--- a/black.py
+++ b/black.py
@@ -3,7 +3,7 @@ from asyncio.base_events import BaseEventLoop
 from concurrent.futures import Executor, ProcessPoolExecutor
 from datetime import datetime
 from enum import Enum, Flag
 from concurrent.futures import Executor, ProcessPoolExecutor
 from datetime import datetime
 from enum import Enum, Flag

-from functools import partial, wraps
+from functools import lru_cache, partial, wraps

 import io
 import keyword
 import logging
 import io
 import keyword
 import logging


@@ -38,6 +38,7 @@ from typing import (
 from appdirs import user_cache_dir
 from attr import dataclass, Factory
 import click
 from appdirs import user_cache_dir
 from attr import dataclass, Factory
 import click

+import toml

 
 # lib2to3 fork
 from blib2to3.pytree import Node, Leaf, type_repr
 
 # lib2to3 fork
 from blib2to3.pytree import Node, Leaf, type_repr


@@ -46,7 +47,7 @@ from blib2to3.pgen2 import driver, token
 from blib2to3.pgen2.parse import ParseError
 
 
 from blib2to3.pgen2.parse import ParseError
 
 

-__version__ = "18.5b1"
+__version__ = "18.6b2"

 DEFAULT_LINE_LENGTH = 88
 DEFAULT_EXCLUDES = (
     r"/(\.git|\.hg|\.mypy_cache|\.tox|\.venv|_build|buck-out|build|dist)/"
 DEFAULT_LINE_LENGTH = 88
 DEFAULT_EXCLUDES = (
     r"/(\.git|\.hg|\.mypy_cache|\.tox|\.venv|_build|buck-out|build|dist)/"


@@ -156,7 +157,41 @@ class FileMode(Flag):
         return mode
 
 
         return mode
 
 

-@click.command()
+def read_pyproject_toml(
+    ctx: click.Context, param: click.Parameter, value: Union[str, int, bool, None]
+) -> Optional[str]:
+    """Inject Black configuration from "pyproject.toml" into defaults in `ctx`.
+
+    Returns the path to a successfully found and read configuration file, None
+    otherwise.
+    """
+    assert not isinstance(value, (int, bool)), "Invalid parameter type passed"
+    if not value:
+        root = find_project_root(ctx.params.get("src", ()))
+        path = root / "pyproject.toml"
+        if path.is_file():
+            value = str(path)
+        else:
+            return None
+
+    try:
+        pyproject_toml = toml.load(value)
+        config = pyproject_toml.get("tool", {}).get("black", {})
+    except (toml.TomlDecodeError, OSError) as e:
+        raise click.BadOptionUsage(f"Error reading configuration file: {e}", ctx)
+
+    if not config:
+        return None
+
+    if ctx.default_map is None:
+        ctx.default_map = {}
+    ctx.default_map.update(  # type: ignore  # bad types in .pyi
+        {k.replace("--", "").replace("-", "_"): v for k, v in config.items()}
+    )
+    return value
+
+
+@click.command(context_settings=dict(help_option_names=["-h", "--help"]))

 @click.option(
     "-l",
     "--line-length",
 @click.option(
     "-l",
     "--line-length",


@@ -257,6 +292,16 @@ class FileMode(Flag):
     type=click.Path(
         exists=True, file_okay=True, dir_okay=True, readable=True, allow_dash=True
     ),
     type=click.Path(
         exists=True, file_okay=True, dir_okay=True, readable=True, allow_dash=True
     ),

+    is_eager=True,
+)
+@click.option(
+    "--config",
+    type=click.Path(
+        exists=False, file_okay=True, dir_okay=False, readable=True, allow_dash=False
+    ),
+    is_eager=True,
+    callback=read_pyproject_toml,
+    help="Read configuration from PATH.",

 )
 @click.pass_context
 def main(
 )
 @click.pass_context
 def main(


@@ -272,26 +317,29 @@ def main(
     verbose: bool,
     include: str,
     exclude: str,
     verbose: bool,
     include: str,
     exclude: str,

-    src: List[str],
+    src: Tuple[str],
+    config: Optional[str],

 ) -> None:
     """The uncompromising code formatter."""
     write_back = WriteBack.from_configuration(check=check, diff=diff)
     mode = FileMode.from_configuration(
         py36=py36, pyi=pyi, skip_string_normalization=skip_string_normalization
     )
 ) -> None:
     """The uncompromising code formatter."""
     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()
+    if config and verbose:
+        out(f"Using configuration from {config}.", bold=False, fg="blue")

     try:
     try:

-        include_regex = re.compile(include)
+        include_regex = re_compile_maybe_verbose(include)

     except re.error:
         err(f"Invalid regular expression for include given: {include!r}")
         ctx.exit(2)
     try:
     except re.error:
         err(f"Invalid regular expression for include given: {include!r}")
         ctx.exit(2)
     try:

-        exclude_regex = re.compile(exclude)
+        exclude_regex = re_compile_maybe_verbose(exclude)

     except re.error:
         err(f"Invalid regular expression for exclude given: {exclude!r}")
         ctx.exit(2)
     except re.error:
         err(f"Invalid regular expression for exclude given: {exclude!r}")
         ctx.exit(2)

+    report = Report(check=check, quiet=quiet, verbose=verbose)

     root = find_project_root(src)
     root = find_project_root(src)

+    sources: Set[Path] = set()

     for s in src:
         p = Path(s)
         if p.is_dir():
     for s in src:
         p = Path(s)
         if p.is_dir():


@@ -307,9 +355,8 @@ def main(
         if verbose or not quiet:
             out("No paths given. Nothing to do 😴")
         ctx.exit(0)
         if verbose or not quiet:
             out("No paths given. Nothing to do 😴")
         ctx.exit(0)

-        return

 
 

-    elif len(sources) == 1:
+    if len(sources) == 1:

         reformat_one(
             src=sources.pop(),
             line_length=line_length,
         reformat_one(
             src=sources.pop(),
             line_length=line_length,


@@ -337,8 +384,9 @@ def main(
         finally:
             shutdown(loop)
     if verbose or not quiet:
         finally:
             shutdown(loop)
     if verbose or not quiet:

-        out("All done! ✨ 🍰 ✨")
-        click.echo(str(report))
+        bang = "💥 💔 💥" if report.return_code else "✨ 🍰 ✨"
+        out(f"All done! {bang}")
+        click.secho(str(report), err=True)

     ctx.exit(report.return_code)
 
 
     ctx.exit(report.return_code)
 
 


@@ -1094,6 +1142,13 @@ class Line:
 
         return False
 
 
         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 (
     def maybe_remove_trailing_comma(self, closing: Leaf) -> bool:
         """Remove trailing comma if there is one and it's safe."""
         if not (


@@ -1175,6 +1230,9 @@ class Line:
 
         Provide a non-negative leaf `_index` to speed up the function.
         """
 
         Provide a non-negative leaf `_index` to speed up the function.
         """

+        if not self.comments:
+            return
+

         if _index == -1:
             for _index, _leaf in enumerate(self.leaves):
                 if leaf is _leaf:
         if _index == -1:
             for _index, _leaf in enumerate(self.leaves):
                 if leaf is _leaf:


@@ -1198,18 +1256,18 @@ class Line:
 
     def is_complex_subscript(self, leaf: Leaf) -> bool:
         """Return True iff `leaf` is part of a slice with non-trivial exprs."""
 
     def is_complex_subscript(self, leaf: Leaf) -> bool:
         """Return True iff `leaf` is part of a slice with non-trivial exprs."""

-        open_lsqb = (
-            leaf if leaf.type == token.LSQB else self.bracket_tracker.get_open_lsqb()
-        )
+        open_lsqb = self.bracket_tracker.get_open_lsqb()

         if open_lsqb is None:
             return False
 
         subscript_start = open_lsqb.next_sibling
         if open_lsqb is None:
             return False
 
         subscript_start = open_lsqb.next_sibling

-        if (
-            isinstance(subscript_start, Node)
-            and subscript_start.type == syms.subscriptlist
-        ):
-            subscript_start = child_towards(subscript_start, leaf)
+
+        if isinstance(subscript_start, Node):
+            if subscript_start.type == syms.listmaker:
+                return False
+
+            if 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()
         )


@@ -1827,7 +1885,7 @@ def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str:  # noqa C901
             elif prevp.type == token.EQUAL and prevp_parent.type == syms.argument:
                 return NO
 
             elif prevp.type == token.EQUAL and prevp_parent.type == syms.argument:
                 return NO
 

-        elif t == token.NAME or t == token.NUMBER:
+        elif t in {token.NAME, token.NUMBER, token.STRING}:

             return NO
 
     elif p.type == syms.import_from:
             return NO
 
     elif p.type == syms.import_from:


@@ -2225,24 +2283,35 @@ def right_hand_split(
         # the closing bracket is an optional paren
         and closing_bracket.type == token.RPAR
         and not closing_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
         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}
     ):
         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:
-                if len(body.leaves) == 1 and not is_line_short_enough(
-                    body, line_length=line_length
-                ):
-                    raise CannotSplit(
-                        "Splitting failed, body is still too long and can't be split."
-                    )
+        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 because the head or the tail "
+                    "contains multiline strings which by definition never fit one "
+                    "line."
+                )

 
     ensure_visible(opening_bracket)
     ensure_visible(closing_bracket)
 
     ensure_visible(opening_bracket)
     ensure_visible(closing_bracket)


@@ -2460,8 +2529,8 @@ def normalize_string_quotes(leaf: Leaf) -> None:
 
     prefix = leaf.value[:first_quote_pos]
     unescaped_new_quote = re.compile(rf"(([^\\]|^)(\\\\)*){new_quote}")
 
     prefix = leaf.value[:first_quote_pos]
     unescaped_new_quote = re.compile(rf"(([^\\]|^)(\\\\)*){new_quote}")

-    escaped_new_quote = re.compile(rf"([^\\]|^)\\(\\\\)*{new_quote}")
-    escaped_orig_quote = re.compile(rf"([^\\]|^)\\(\\\\)*{orig_quote}")
+    escaped_new_quote = re.compile(rf"([^\\]|^)\\((?:\\\\)*){new_quote}")
+    escaped_orig_quote = re.compile(rf"([^\\]|^)\\((?:\\\\)*){orig_quote}")

     body = leaf.value[first_quote_pos + len(orig_quote) : -len(orig_quote)]
     if "r" in prefix.casefold():
         if unescaped_new_quote.search(body):
     body = leaf.value[first_quote_pos + len(orig_quote) : -len(orig_quote)]
     if "r" in prefix.casefold():
         if unescaped_new_quote.search(body):


@@ -2472,15 +2541,21 @@ def normalize_string_quotes(leaf: Leaf) -> None:
         # Do not introduce or remove backslashes in raw strings
         new_body = body
     else:
         # Do not introduce or remove backslashes in raw strings
         new_body = body
     else:

-        # remove unnecessary quotes
+        # remove unnecessary escapes

         new_body = sub_twice(escaped_new_quote, rf"\1\2{new_quote}", body)
         if body != new_body:
         new_body = sub_twice(escaped_new_quote, rf"\1\2{new_quote}", body)
         if body != new_body:

-            # Consider the string without unnecessary quotes as the original
+            # Consider the string without unnecessary escapes as the original

             body = new_body
             leaf.value = f"{prefix}{orig_quote}{body}{orig_quote}"
         new_body = sub_twice(escaped_orig_quote, rf"\1\2{orig_quote}", new_body)
         new_body = sub_twice(unescaped_new_quote, rf"\1\\{new_quote}", new_body)
             body = new_body
             leaf.value = f"{prefix}{orig_quote}{body}{orig_quote}"
         new_body = sub_twice(escaped_orig_quote, rf"\1\2{orig_quote}", new_body)
         new_body = sub_twice(unescaped_new_quote, rf"\1\\{new_quote}", new_body)

-    if new_quote == '"""' and new_body[-1] == '"':
+    if "f" in prefix.casefold():
+        matches = re.findall(r"[^{]\{(.*?)\}[^}]", new_body)
+        for m in matches:
+            if "\\" in str(m):
+                # Do not introduce backslashes in interpolated expressions
+                return
+    if new_quote == '"""' and new_body[-1:] == '"':

         # edge case:
         new_body = new_body[:-1] + '\\"'
     orig_escape_count = body.count("\\")
         # edge case:
         new_body = new_body[:-1] + '\\"'
     orig_escape_count = body.count("\\")


@@ -2543,7 +2618,7 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
 
 
 def maybe_make_parens_invisible_in_atom(node: LN) -> bool:
 
 
 def maybe_make_parens_invisible_in_atom(node: LN) -> bool:

-    """If it's safe, make the parens in the atom `node` invisible, recusively."""
+    """If it's safe, make the parens in the atom `node` invisible, recursively."""

     if (
         node.type != syms.atom
         or is_empty_tuple(node)
     if (
         node.type != syms.atom
         or is_empty_tuple(node)


@@ -2866,16 +2941,29 @@ def gen_python_files_in_dir(
     """Generate all files under `path` whose paths are not excluded by the
     `exclude` regex, but are included by the `include` regex.
 
     """Generate all files under `path` whose paths are not excluded by the
     `exclude` regex, but are included by the `include` regex.
 

+    Symbolic links pointing outside of the root directory are ignored.
+

     `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():
     `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()
+        try:
+            normalized_path = "/" + child.resolve().relative_to(root).as_posix()
+        except ValueError:
+            if child.is_symlink():
+                report.path_ignored(
+                    child,
+                    "is a symbolic link that points outside of the root directory",
+                )
+                continue
+
+            raise
+

         if child.is_dir():
             normalized_path += "/"
         exclude_match = exclude.search(normalized_path)
         if exclude_match and exclude_match.group(0):
         if child.is_dir():
             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}")
+            report.path_ignored(child, f"matches the --exclude regular expression")

             continue
 
         if child.is_dir():
             continue
 
         if child.is_dir():


@@ -2887,7 +2975,8 @@ def gen_python_files_in_dir(
                 yield child
 
 
                 yield child
 
 

-def find_project_root(srcs: List[str]) -> Path:
+@lru_cache()
+def find_project_root(srcs: Iterable[str]) -> Path:

     """Return a directory containing .git, .hg, or pyproject.toml.
 
     That directory can be one of the directories passed in `srcs` or their
     """Return a directory containing .git, .hg, or pyproject.toml.
 
     That directory can be one of the directories passed in `srcs` or their


@@ -3145,6 +3234,16 @@ def sub_twice(regex: Pattern[str], replacement: str, original: str) -> str:
     return regex.sub(replacement, regex.sub(replacement, original))
 
 
     return regex.sub(replacement, regex.sub(replacement, original))
 
 

+def re_compile_maybe_verbose(regex: str) -> Pattern[str]:
+    """Compile a regular expression string in `regex`.
+
+    If it contains newlines, use verbose mode.
+    """
+    if "\n" in regex:
+        regex = "(?x)" + regex
+    return re.compile(regex)
+
+

 def enumerate_reversed(sequence: Sequence[T]) -> Iterator[Tuple[Index, T]]:
     """Like `reversed(enumerate(sequence))` if that were possible."""
     index = len(sequence) - 1
 def enumerate_reversed(sequence: Sequence[T]) -> Iterator[Tuple[Index, T]]:
     """Like `reversed(enumerate(sequence))` if that were possible."""
     index = len(sequence) - 1


@@ -3190,6 +3289,42 @@ 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?
 
 def can_omit_invisible_parens(line: Line, line_length: int) -> bool:
     """Does `line` have a shape safe to reformat without optional parens around it?
 


@@ -3280,12 +3415,7 @@ def can_omit_invisible_parens(line: Line, line_length: int) -> bool:
 
 
 def get_cache_file(line_length: int, mode: FileMode) -> Path:
 
 
 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"
-    )
+    return CACHE_DIR / f"cache.{line_length}.{mode.value}.pickle"

 
 
 def read_cache(line_length: int, mode: FileMode) -> Cache:
 
 
 def read_cache(line_length: int, mode: FileMode) -> Cache: