Use early returns in `get_grammars()` to save an indentation level

[etc/vim.git] / black.py

diff --git a/black.py b/black.py
index 40b4d2b12876013b0925c9a1c706110befeb9b28..953d532a6567fa3493f1b454df79452b2e831b93 100644 (file)
--- a/black.py
+++ b/black.py
@@ -43,6 +43,7 @@ from attr import dataclass, evolve, Factory
 import click
 import toml
 from typed_ast import ast3, ast27
 import click
 import toml
 from typed_ast import ast3, ast27

+from pathspec import PathSpec

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


@@ -51,12 +52,10 @@ from blib2to3.pgen2 import driver, token
 from blib2to3.pgen2.grammar import Grammar
 from blib2to3.pgen2.parse import ParseError
 
 from blib2to3.pgen2.grammar import Grammar
 from blib2to3.pgen2.parse import ParseError
 

-from _version import version as __version__
+from _black_version import version as __version__

 
 DEFAULT_LINE_LENGTH = 88
 
 DEFAULT_LINE_LENGTH = 88

-DEFAULT_EXCLUDES = (
-    r"/(\.eggs|\.git|\.hg|\.mypy_cache|\.nox|\.tox|\.venv|_build|buck-out|build|dist)/"
-)
+DEFAULT_EXCLUDES = r"/(\.eggs|\.git|\.hg|\.mypy_cache|\.nox|\.tox|\.venv|\.svn|_build|buck-out|build|dist)/"  # noqa: B950

 DEFAULT_INCLUDES = r"\.pyi?$"
 CACHE_DIR = Path(user_cache_dir("black", version=__version__))
 
 DEFAULT_INCLUDES = r"\.pyi?$"
 CACHE_DIR = Path(user_cache_dir("black", version=__version__))
 


@@ -438,7 +437,9 @@ def main(
         p = Path(s)
         if p.is_dir():
             sources.update(
         p = Path(s)
         if p.is_dir():
             sources.update(

-                gen_python_files_in_dir(p, root, include_regex, exclude_regex, report)
+                gen_python_files_in_dir(
+                    p, root, include_regex, exclude_regex, report, get_gitignore(root)
+                )

             )
         elif p.is_file() or s == "-":
             # if a file was explicitly given, we don't care about its extension
             )
         elif p.is_file() or s == "-":
             # if a file was explicitly given, we don't care about its extension


@@ -786,7 +787,8 @@ def get_grammars(target_versions: Set[TargetVersion]) -> List[Grammar]:
             # Python 2.7
             pygram.python_grammar,
         ]
             # Python 2.7
             pygram.python_grammar,
         ]

-    elif all(version.is_python2() for version in target_versions):
+
+    if all(version.is_python2() for version in target_versions):

         # Python 2-only code, so try Python 2 grammars.
         return [
             # Python 2.7 with future print_function import
         # Python 2-only code, so try Python 2 grammars.
         return [
             # Python 2.7 with future print_function import


@@ -794,21 +796,21 @@ def get_grammars(target_versions: Set[TargetVersion]) -> List[Grammar]:
             # Python 2.7
             pygram.python_grammar,
         ]
             # Python 2.7
             pygram.python_grammar,
         ]

-    else:
-        # Python 3-compatible code, so only try Python 3 grammar.
-        grammars = []
-        # If we have to parse both, try to parse async as a keyword first
-        if not supports_feature(target_versions, Feature.ASYNC_IDENTIFIERS):
-            # Python 3.7+
-            grammars.append(
-                pygram.python_grammar_no_print_statement_no_exec_statement_async_keywords  # noqa: B950
-            )
-        if not supports_feature(target_versions, Feature.ASYNC_KEYWORDS):
-            # Python 3.0-3.6
-            grammars.append(pygram.python_grammar_no_print_statement_no_exec_statement)
-        # At least one of the above branches must have been taken, because every Python
-        # version has exactly one of the two 'ASYNC_*' flags
-        return grammars
+
+    # Python 3-compatible code, so only try Python 3 grammar.
+    grammars = []
+    # If we have to parse both, try to parse async as a keyword first
+    if not supports_feature(target_versions, Feature.ASYNC_IDENTIFIERS):
+        # Python 3.7+
+        grammars.append(
+            pygram.python_grammar_no_print_statement_no_exec_statement_async_keywords
+        )
+    if not supports_feature(target_versions, Feature.ASYNC_KEYWORDS):
+        # Python 3.0-3.6
+        grammars.append(pygram.python_grammar_no_print_statement_no_exec_statement)
+    # At least one of the above branches must have been taken, because every Python
+    # version has exactly one of the two 'ASYNC_*' flags
+    return grammars

 
 
 def lib2to3_parse(src_txt: str, target_versions: Iterable[TargetVersion] = ()) -> Node:
 
 
 def lib2to3_parse(src_txt: str, target_versions: Iterable[TargetVersion] = ()) -> Node:


@@ -1716,39 +1718,6 @@ class LineGenerator(Visitor[Line]):
                 self.current_line.append(node)
         yield from super().visit_default(node)
 
                 self.current_line.append(node)
         yield from super().visit_default(node)
 

-    def visit_atom(self, node: Node) -> Iterator[Line]:
-        # Always make parentheses invisible around a single node, because it should
-        # not be needed (except in the case of yield, where removing the parentheses
-        # produces a SyntaxError).
-        if (
-            len(node.children) == 3
-            and isinstance(node.children[0], Leaf)
-            and node.children[0].type == token.LPAR
-            and isinstance(node.children[2], Leaf)
-            and node.children[2].type == token.RPAR
-            and isinstance(node.children[1], Leaf)
-            and not (
-                node.children[1].type == token.NAME
-                and node.children[1].value == "yield"
-            )
-        ):
-            node.children[0].value = ""
-            node.children[2].value = ""
-        yield from super().visit_default(node)
-
-    def visit_factor(self, node: Node) -> Iterator[Line]:
-        """Force parentheses between a unary op and a binary power:
-
-        -2 ** 8 -> -(2 ** 8)
-        """
-        child = node.children[1]
-        if child.type == syms.power and len(child.children) == 3:
-            lpar = Leaf(token.LPAR, "(")
-            rpar = Leaf(token.RPAR, ")")
-            index = child.remove() or 0
-            node.insert_child(index, Node(syms.atom, [lpar, child, rpar]))
-        yield from self.visit_default(node)
-

     def visit_INDENT(self, node: Node) -> Iterator[Line]:
         """Increase indentation level, maybe yield a line."""
         # In blib2to3 INDENT never holds comments.
     def visit_INDENT(self, node: Node) -> Iterator[Line]:
         """Increase indentation level, maybe yield a line."""
         # In blib2to3 INDENT never holds comments.


@@ -1848,6 +1817,23 @@ class LineGenerator(Visitor[Line]):
             yield from self.line()
         yield from self.visit_default(leaf)
 
             yield from self.line()
         yield from self.visit_default(leaf)
 

+    def visit_factor(self, node: Node) -> Iterator[Line]:
+        """Force parentheses between a unary op and a binary power:
+
+        -2 ** 8 -> -(2 ** 8)
+        """
+        _operator, operand = node.children
+        if (
+            operand.type == syms.power
+            and len(operand.children) == 3
+            and operand.children[1].type == token.DOUBLESTAR
+        ):
+            lpar = Leaf(token.LPAR, "(")
+            rpar = Leaf(token.RPAR, ")")
+            index = operand.remove() or 0
+            node.insert_child(index, Node(syms.atom, [lpar, operand, rpar]))
+        yield from self.visit_default(node)
+

     def __attrs_post_init__(self) -> None:
         """You are in a twisty little maze of passages."""
         v = self.visit_stmt
     def __attrs_post_init__(self) -> None:
         """You are in a twisty little maze of passages."""
         v = self.visit_stmt


@@ -2421,7 +2407,9 @@ def split_line(
             # All splits failed, best effort split with no omits.
             # This mostly happens to multiline strings that are by definition
             # reported as not fitting a single line.
             # All splits failed, best effort split with no omits.
             # This mostly happens to multiline strings that are by definition
             # reported as not fitting a single line.

-            yield from right_hand_split(line, line_length, features=features)
+            # line_length=1 here was historically a bug that somehow became a feature.
+            # See #762 and #781 for the full story.
+            yield from right_hand_split(line, line_length=1, features=features)

 
         if line.inside_brackets:
             split_funcs = [delimiter_split, standalone_comment_split, rhs]
 
         if line.inside_brackets:
             split_funcs = [delimiter_split, standalone_comment_split, rhs]


@@ -2620,9 +2608,11 @@ def bracket_split_build_line(
             # Since body is a new indent level, remove spurious leading whitespace.
             normalize_prefix(leaves[0], inside_brackets=True)
             # Ensure a trailing comma for imports and standalone function arguments, but
             # Since body is a new indent level, remove spurious leading whitespace.
             normalize_prefix(leaves[0], inside_brackets=True)
             # Ensure a trailing comma for imports and standalone function arguments, but

-            # be careful not to add one after any comments.
-            no_commas = original.is_def and not any(
-                l.type == token.COMMA for l in leaves
+            # be careful not to add one after any comments or within type annotations.
+            no_commas = (
+                original.is_def
+                and opening_bracket.value == "("
+                and not any(l.type == token.COMMA for l in leaves)

             )
 
             if original.is_import or no_commas:
             )
 
             if original.is_import or no_commas:


@@ -2779,9 +2769,7 @@ def is_type_comment(leaf: Leaf, suffix: str = "") -> bool:
     Only returns true for type comments for now."""
     t = leaf.type
     v = leaf.value
     Only returns true for type comments for now."""
     t = leaf.type
     v = leaf.value

-    return t in {token.COMMENT, t == STANDALONE_COMMENT} and v.startswith(
-        "# type:" + suffix
-    )
+    return t in {token.COMMENT, STANDALONE_COMMENT} and v.startswith("# type:" + suffix)

 
 
 def normalize_prefix(leaf: Leaf, *, inside_brackets: bool) -> None:
 
 
 def normalize_prefix(leaf: Leaf, *, inside_brackets: bool) -> None:


@@ -3477,12 +3465,23 @@ def get_future_imports(node: Node) -> Set[str]:
     return imports
 
 
     return imports
 
 

+@lru_cache()
+def get_gitignore(root: Path) -> PathSpec:
+    """ Return a PathSpec matching gitignore content if present."""
+    gitignore = root / ".gitignore"
+    if not gitignore.is_file():
+        return PathSpec.from_lines("gitwildmatch", [])
+    else:
+        return PathSpec.from_lines("gitwildmatch", gitignore.open())
+
+

 def gen_python_files_in_dir(
     path: Path,
     root: Path,
     include: Pattern[str],
     exclude: Pattern[str],
     report: "Report",
 def gen_python_files_in_dir(
     path: Path,
     root: Path,
     include: Pattern[str],
     exclude: Pattern[str],
     report: "Report",

+    gitignore: PathSpec,

 ) -> Iterator[Path]:
     """Generate all files under `path` whose paths are not excluded by the
     `exclude` regex, but are included by the `include` regex.
 ) -> Iterator[Path]:
     """Generate all files under `path` whose paths are not excluded by the
     `exclude` regex, but are included by the `include` regex.


@@ -3493,8 +3492,17 @@ def gen_python_files_in_dir(
     """
     assert root.is_absolute(), f"INTERNAL ERROR: `root` must be absolute but is {root}"
     for child in path.iterdir():
     """
     assert root.is_absolute(), f"INTERNAL ERROR: `root` must be absolute but is {root}"
     for child in path.iterdir():

+        # First ignore files matching .gitignore
+        if gitignore.match_file(child.as_posix()):
+            report.path_ignored(child, f"matches the .gitignore file content")
+            continue
+
+        # Then ignore with `exclude` option.

         try:
             normalized_path = "/" + child.resolve().relative_to(root).as_posix()
         try:
             normalized_path = "/" + child.resolve().relative_to(root).as_posix()

+        except OSError as e:
+            report.path_ignored(child, f"cannot be read because {e}")
+            continue

         except ValueError:
             if child.is_symlink():
                 report.path_ignored(
         except ValueError:
             if child.is_symlink():
                 report.path_ignored(


@@ -3506,13 +3514,16 @@ def gen_python_files_in_dir(
 
         if child.is_dir():
             normalized_path += "/"
 
         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 the --exclude regular expression")
             continue
 
         if child.is_dir():
         exclude_match = exclude.search(normalized_path)
         if exclude_match and exclude_match.group(0):
             report.path_ignored(child, f"matches the --exclude regular expression")
             continue
 
         if child.is_dir():

-            yield from gen_python_files_in_dir(child, root, include, exclude, report)
+            yield from gen_python_files_in_dir(
+                child, root, include, exclude, report, gitignore
+            )

 
         elif child.is_file():
             include_match = include.search(normalized_path)
 
         elif child.is_file():
             include_match = include.search(normalized_path)


@@ -3776,8 +3787,10 @@ def dump_to_file(*output: str) -> str:
 
 @contextmanager
 def nullcontext() -> Iterator[None]:
 
 @contextmanager
 def nullcontext() -> Iterator[None]:

-    """Return context manager that does nothing.
-    Similar to `nullcontext` from python 3.7"""
+    """Return an empty context manager.
+
+    To be used like `nullcontext` in Python 3.7.
+    """

     yield
 
 
     yield
 
 


@@ -4030,7 +4043,7 @@ def read_cache(mode: FileMode) -> Cache:
     with cache_file.open("rb") as fobj:
         try:
             cache: Cache = pickle.load(fobj)
     with cache_file.open("rb") as fobj:
         try:
             cache: Cache = pickle.load(fobj)

-        except pickle.UnpicklingError:
+        except (pickle.UnpicklingError, ValueError):

             return {}
 
     return cache
             return {}
 
     return cache


@@ -4065,7 +4078,7 @@ def write_cache(cache: Cache, sources: Iterable[Path], mode: FileMode) -> None:
         CACHE_DIR.mkdir(parents=True, exist_ok=True)
         new_cache = {**cache, **{src.resolve(): get_cache_info(src) for src in sources}}
         with tempfile.NamedTemporaryFile(dir=str(cache_file.parent), delete=False) as f:
         CACHE_DIR.mkdir(parents=True, exist_ok=True)
         new_cache = {**cache, **{src.resolve(): get_cache_info(src) for src in sources}}
         with tempfile.NamedTemporaryFile(dir=str(cache_file.parent), delete=False) as f:

-            pickle.dump(new_cache, f, protocol=pickle.HIGHEST_PROTOCOL)
+            pickle.dump(new_cache, f, protocol=4)

         os.replace(f.name, cache_file)
     except OSError:
         pass
         os.replace(f.name, cache_file)
     except OSError:
         pass