]> git.madduck.net Git - etc/vim.git/blobdiff - black.py

madduck's git repository

Every one of the projects in this repository is available at the canonical URL git://git.madduck.net/madduck/pub/<projectpath> — see each project's metadata for the exact URL.

All patches and comments are welcome. Please squash your changes to logical commits before using git-format-patch and git-send-email to patches@git.madduck.net. If you'd read over the Git project's submission guidelines and adhered to them, I'd be especially grateful.

SSH access, as well as push access can be individually arranged.

If you use my repositories frequently, consider adding the following snippet to ~/.gitconfig and using the third clone URL listed for each project:

[url "git://git.madduck.net/madduck/"]
  insteadOf = madduck:

show how to exclude individual files in the exclude example (#663)
[etc/vim.git] / black.py
index 522474123092e46c6d3d3bf26b7fe0211141d634..c4b91328e7001187fdbca136e2c3c2a0f9ef16d6 100644 (file)
--- a/black.py
+++ b/black.py
@@ -5,15 +5,16 @@ from datetime import datetime
 from enum import Enum, Flag
 from functools import lru_cache, partial, wraps
 import io
-import keyword
+import itertools
 import logging
-from multiprocessing import Manager
+from multiprocessing import Manager, freeze_support
 import os
 from pathlib import Path
 import pickle
 import re
 import signal
 import sys
+import tempfile
 import tokenize
 from typing import (
     Any,
@@ -47,10 +48,10 @@ from blib2to3.pgen2 import driver, token
 from blib2to3.pgen2.parse import ParseError
 
 
-__version__ = "18.6b4"
+__version__ = "18.9b0"
 DEFAULT_LINE_LENGTH = 88
 DEFAULT_EXCLUDES = (
-    r"/(\.git|\.hg|\.mypy_cache|\.tox|\.venv|_build|buck-out|build|dist)/"
+    r"/(\.eggs|\.git|\.hg|\.mypy_cache|\.nox|\.tox|\.venv|_build|buck-out|build|dist)/"
 )
 DEFAULT_INCLUDES = r"\.pyi?$"
 CACHE_DIR = Path(user_cache_dir("black", version=__version__))
@@ -115,10 +116,16 @@ class FileMode(Flag):
     PYTHON36 = 1
     PYI = 2
     NO_STRING_NORMALIZATION = 4
+    NO_NUMERIC_UNDERSCORE_NORMALIZATION = 8
 
     @classmethod
     def from_configuration(
-        cls, *, py36: bool, pyi: bool, skip_string_normalization: bool
+        cls,
+        *,
+        py36: bool,
+        pyi: bool,
+        skip_string_normalization: bool,
+        skip_numeric_underscore_normalization: bool,
     ) -> "FileMode":
         mode = cls.AUTO_DETECT
         if py36:
@@ -127,6 +134,8 @@ class FileMode(Flag):
             mode |= cls.PYI
         if skip_string_normalization:
             mode |= cls.NO_STRING_NORMALIZATION
+        if skip_numeric_underscore_normalization:
+            mode |= cls.NO_NUMERIC_UNDERSCORE_NORMALIZATION
         return mode
 
 
@@ -151,7 +160,9 @@ def read_pyproject_toml(
         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)
+        raise click.FileError(
+            filename=value, hint=f"Error reading configuration file: {e}"
+        )
 
     if not config:
         return None
@@ -196,6 +207,12 @@ def read_pyproject_toml(
     is_flag=True,
     help="Don't normalize string quotes or prefixes.",
 )
+@click.option(
+    "-N",
+    "--skip-numeric-underscore-normalization",
+    is_flag=True,
+    help="Don't normalize underscores in numeric literals.",
+)
 @click.option(
     "--check",
     is_flag=True,
@@ -286,6 +303,7 @@ def main(
     pyi: bool,
     py36: bool,
     skip_string_normalization: bool,
+    skip_numeric_underscore_normalization: bool,
     quiet: bool,
     verbose: bool,
     include: str,
@@ -296,7 +314,10 @@ def main(
     """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
+        py36=py36,
+        pyi=pyi,
+        skip_string_normalization=skip_string_normalization,
+        skip_numeric_underscore_normalization=skip_numeric_underscore_normalization,
     )
     if config and verbose:
         out(f"Using configuration from {config}.", bold=False, fg="blue")
@@ -607,7 +628,7 @@ def format_str(
 
     `line_length` determines how many characters per line are allowed.
     """
-    src_node = lib2to3_parse(src_contents)
+    src_node = lib2to3_parse(src_contents.lstrip())
     dst_contents = ""
     future_imports = get_future_imports(src_node)
     is_pyi = bool(mode & FileMode.PYI)
@@ -618,7 +639,8 @@ def format_str(
         remove_u_prefix=py36 or "unicode_literals" in future_imports,
         is_pyi=is_pyi,
         normalize_strings=normalize_strings,
-        allow_underscores=py36,
+        allow_underscores=py36
+        and not bool(mode & FileMode.NO_NUMERIC_UNDERSCORE_NORMALIZATION),
     )
     elt = EmptyLineTracker(is_pyi=is_pyi)
     empty_line = Line()
@@ -660,7 +682,6 @@ 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":
         src_txt += "\n"
     for grammar in GRAMMARS:
@@ -756,9 +777,7 @@ class DebugVisitor(Visitor[T]):
         list(v.visit(code))
 
 
-KEYWORDS = set(keyword.kwlist)
 WHITESPACE = {token.DEDENT, token.INDENT, token.NEWLINE}
-FLOW_CONTROL = {"return", "raise", "break", "continue"}
 STATEMENT = {
     syms.if_stmt,
     syms.while_stmt,
@@ -1007,7 +1026,9 @@ class Line:
 
     depth: int = 0
     leaves: List[Leaf] = Factory(list)
-    comments: List[Tuple[Index, Leaf]] = Factory(list)
+    # The LeafID keys of comments must remain ordered by the corresponding leaf's index
+    # in leaves
+    comments: Dict[LeafID, List[Leaf]] = Factory(dict)
     bracket_tracker: BracketTracker = Factory(BracketTracker)
     inside_brackets: bool = False
     should_explode: bool = False
@@ -1214,43 +1235,35 @@ class Line:
         if comment.type != token.COMMENT:
             return False
 
-        after = len(self.leaves) - 1
-        if after == -1:
+        if not self.leaves:
             comment.type = STANDALONE_COMMENT
             comment.prefix = ""
             return False
 
         else:
-            self.comments.append((after, comment))
-            return True
-
-    def comments_after(self, leaf: Leaf, _index: int = -1) -> Iterator[Leaf]:
-        """Generate comments that should appear directly after `leaf`.
-
-        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:
-                    break
-
+            leaf_id = id(self.leaves[-1])
+            if leaf_id not in self.comments:
+                self.comments[leaf_id] = [comment]
             else:
-                return
+                self.comments[leaf_id].append(comment)
+            return True
 
-        for index, comment_after in self.comments:
-            if _index == index:
-                yield comment_after
+    def comments_after(self, leaf: Leaf) -> List[Leaf]:
+        """Generate comments that should appear directly after `leaf`."""
+        return self.comments.get(id(leaf), [])
 
     def remove_trailing_comma(self) -> None:
         """Remove the trailing comma and moves the comments attached to it."""
-        comma_index = len(self.leaves) - 1
-        for i in range(len(self.comments)):
-            comment_index, comment = self.comments[i]
-            if comment_index == comma_index:
-                self.comments[i] = (comma_index - 1, comment)
+        # Remember, the LeafID keys of self.comments are ordered by the
+        # corresponding leaf's index in self.leaves
+        # If id(self.leaves[-2]) is in self.comments, the order doesn't change.
+        # Otherwise, we insert it into self.comments, and it becomes the last entry.
+        # However, since we delete id(self.leaves[-1]) from self.comments, the invariant
+        # is maintained
+        self.comments.setdefault(id(self.leaves[-2]), []).extend(
+            self.comments.get(id(self.leaves[-1]), [])
+        )
+        self.comments.pop(id(self.leaves[-1]), None)
         self.leaves.pop()
 
     def is_complex_subscript(self, leaf: Leaf) -> bool:
@@ -1282,7 +1295,7 @@ class Line:
         res = f"{first.prefix}{indent}{first.value}"
         for leaf in leaves:
             res += str(leaf)
-        for _, comment in self.comments:
+        for comment in itertools.chain.from_iterable(self.comments.values()):
             res += str(comment)
         return res + "\n"
 
@@ -1593,7 +1606,7 @@ BRACKETS = OPENING_BRACKETS | CLOSING_BRACKETS
 ALWAYS_NO_SPACE = CLOSING_BRACKETS | {token.COMMA, STANDALONE_COMMENT}
 
 
-def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str:  # noqa C901
+def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str:  # noqa: C901
     """Return whitespace prefix if needed for the given `leaf`.
 
     `complex_subscript` signals whether the given leaf is part of a subscription
@@ -1874,7 +1887,7 @@ def container_of(leaf: Leaf) -> LN:
     return container
 
 
-def is_split_after_delimiter(leaf: Leaf, previous: Leaf = None) -> int:
+def is_split_after_delimiter(leaf: Leaf, previous: Optional[Leaf] = None) -> int:
     """Return the priority of the `leaf` delimiter, given a line break after it.
 
     The delimiter priorities returned here are from those delimiters that would
@@ -1888,7 +1901,7 @@ def is_split_after_delimiter(leaf: Leaf, previous: Leaf = None) -> int:
     return 0
 
 
-def is_split_before_delimiter(leaf: Leaf, previous: Leaf = None) -> int:
+def is_split_before_delimiter(leaf: Leaf, previous: Optional[Leaf] = None) -> int:
     """Return the priority of the `leaf` delimiter, given a line break before it.
 
     The delimiter priorities returned here are from those delimiters that would
@@ -2063,8 +2076,8 @@ def list_comments(prefix: str, *, is_endmarker: bool) -> List[ProtoComment]:
 def make_comment(content: str) -> str:
     """Return a consistently formatted comment from the given `content` string.
 
-    All comments (except for "##", "#!", "#:") should have a single space between
-    the hash sign and the content.
+    All comments (except for "##", "#!", "#:", '#'", "#%%") should have a single
+    space between the hash sign and the content.
 
     If `content` didn't start with a hash sign, one is provided.
     """
@@ -2074,7 +2087,7 @@ def make_comment(content: str) -> str:
 
     if content[0] == "#":
         content = content[1:]
-    if content and content[0] not in " !:#":
+    if content and content[0] not in " !:#'%":
         content = " " + content
     return "#" + content
 
@@ -2097,8 +2110,19 @@ def split_line(
         return
 
     line_str = str(line).strip("\n")
-    if not line.should_explode and is_line_short_enough(
-        line, line_length=line_length, line_str=line_str
+
+    # we don't want to split special comments like type annotations
+    # https://github.com/python/typing/issues/186
+    has_special_comment = False
+    for leaf in line.leaves:
+        for comment in line.comments_after(leaf):
+            if leaf.type == token.COMMA and is_special_comment(comment):
+                has_special_comment = True
+
+    if (
+        not has_special_comment
+        and not line.should_explode
+        and is_line_short_enough(line, line_length=line_length, line_str=line_str)
     ):
         yield line
         return
@@ -2137,7 +2161,7 @@ def split_line(
                 result.extend(
                     split_line(l, line_length=line_length, inner=True, py36=py36)
                 )
-        except CannotSplit as cs:
+        except CannotSplit:
             continue
 
         else:
@@ -2155,9 +2179,6 @@ def left_hand_split(line: Line, py36: bool = False) -> Iterator[Line]:
     Prefer RHS otherwise.  This is why this function is not symmetrical with
     :func:`right_hand_split` which also handles optional parentheses.
     """
-    head = Line(depth=line.depth)
-    body = Line(depth=line.depth + 1, inside_brackets=True)
-    tail = Line(depth=line.depth)
     tail_leaves: List[Leaf] = []
     body_leaves: List[Leaf] = []
     head_leaves: List[Leaf] = []
@@ -2175,15 +2196,12 @@ def left_hand_split(line: Line, py36: bool = False) -> Iterator[Line]:
             if leaf.type in OPENING_BRACKETS:
                 matching_bracket = leaf
                 current_leaves = body_leaves
-    # Since body is a new indent level, remove spurious leading whitespace.
-    if body_leaves:
-        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):
-        for leaf in leaves:
-            result.append(leaf, preformatted=True)
-            for comment_after in line.comments_after(leaf):
-                result.append(comment_after, preformatted=True)
+    if not matching_bracket:
+        raise CannotSplit("No brackets found")
+
+    head = bracket_split_build_line(head_leaves, line, matching_bracket)
+    body = bracket_split_build_line(body_leaves, line, matching_bracket, is_body=True)
+    tail = bracket_split_build_line(tail_leaves, line, matching_bracket)
     bracket_split_succeeded_or_raise(head, body, tail)
     for result in (head, body, tail):
         if result:
@@ -2201,9 +2219,6 @@ def right_hand_split(
 
     Note: running this function modifies `bracket_depth` on the leaves of `line`.
     """
-    head = Line(depth=line.depth)
-    body = Line(depth=line.depth + 1, inside_brackets=True)
-    tail = Line(depth=line.depth)
     tail_leaves: List[Leaf] = []
     body_leaves: List[Leaf] = []
     head_leaves: List[Leaf] = []
@@ -2220,25 +2235,18 @@ def right_hand_split(
                 opening_bracket = leaf.opening_bracket
                 closing_bracket = leaf
                 current_leaves = body_leaves
-    tail_leaves.reverse()
-    body_leaves.reverse()
-    head_leaves.reverse()
-    # Since body is a new indent level, remove spurious leading whitespace.
-    if body_leaves:
-        normalize_prefix(body_leaves[0], inside_brackets=True)
-    if not head_leaves:
-        # No `head` means the split failed. Either `tail` has all content or
+    if not (opening_bracket and closing_bracket and head_leaves):
+        # If there is no opening or closing_bracket that means the split failed and
+        # all content is in the tail.  Otherwise, if `head_leaves` are empty, it means
         # the matching `opening_bracket` wasn't available on `line` anymore.
         raise CannotSplit("No brackets found")
 
-    # Build the new lines.
-    for result, leaves in (head, head_leaves), (body, body_leaves), (tail, tail_leaves):
-        for leaf in leaves:
-            result.append(leaf, preformatted=True)
-            for comment_after in line.comments_after(leaf):
-                result.append(comment_after, preformatted=True)
-    assert opening_bracket and closing_bracket
-    body.should_explode = should_explode(body, opening_bracket)
+    tail_leaves.reverse()
+    body_leaves.reverse()
+    head_leaves.reverse()
+    head = bracket_split_build_line(head_leaves, line, opening_bracket)
+    body = bracket_split_build_line(body_leaves, line, opening_bracket, is_body=True)
+    tail = bracket_split_build_line(tail_leaves, line, opening_bracket)
     bracket_split_succeeded_or_raise(head, body, tail)
     if (
         # the body shouldn't be exploded
@@ -2312,6 +2320,35 @@ def bracket_split_succeeded_or_raise(head: Line, body: Line, tail: Line) -> None
             )
 
 
+def bracket_split_build_line(
+    leaves: List[Leaf], original: Line, opening_bracket: Leaf, *, is_body: bool = False
+) -> Line:
+    """Return a new line with given `leaves` and respective comments from `original`.
+
+    If `is_body` is True, the result line is one-indented inside brackets and as such
+    has its first leaf's prefix normalized and a trailing comma added when expected.
+    """
+    result = Line(depth=original.depth)
+    if is_body:
+        result.inside_brackets = True
+        result.depth += 1
+        if leaves:
+            # Since body is a new indent level, remove spurious leading whitespace.
+            normalize_prefix(leaves[0], inside_brackets=True)
+            # Ensure a trailing comma when expected.
+            if original.is_import:
+                if leaves[-1].type != token.COMMA:
+                    leaves.append(Leaf(token.COMMA, ","))
+    # Populate the line
+    for leaf in leaves:
+        result.append(leaf, preformatted=True)
+        for comment_after in original.comments_after(leaf):
+            result.append(comment_after, preformatted=True)
+    if is_body:
+        result.should_explode = should_explode(result, opening_bracket)
+    return result
+
+
 def dont_increase_indentation(split_func: SplitFunc) -> SplitFunc:
     """Normalize prefix of the first leaf in every line returned by `split_func`.
 
@@ -2358,16 +2395,16 @@ def delimiter_split(line: Line, py36: bool = False) -> Iterator[Line]:
         nonlocal current_line
         try:
             current_line.append_safe(leaf, preformatted=True)
-        except ValueError as ve:
+        except ValueError:
             yield current_line
 
             current_line = Line(depth=line.depth, inside_brackets=line.inside_brackets)
             current_line.append(leaf)
 
-    for index, leaf in enumerate(line.leaves):
+    for leaf in line.leaves:
         yield from append_to_line(leaf)
 
-        for comment_after in line.comments_after(leaf, index):
+        for comment_after in line.comments_after(leaf):
             yield from append_to_line(comment_after)
 
         lowest_depth = min(lowest_depth, leaf.bracket_depth)
@@ -2404,16 +2441,16 @@ def standalone_comment_split(line: Line, py36: bool = False) -> Iterator[Line]:
         nonlocal current_line
         try:
             current_line.append_safe(leaf, preformatted=True)
-        except ValueError as ve:
+        except ValueError:
             yield current_line
 
             current_line = Line(depth=line.depth, inside_brackets=line.inside_brackets)
             current_line.append(leaf)
 
-    for index, leaf in enumerate(line.leaves):
+    for leaf in line.leaves:
         yield from append_to_line(leaf)
 
-        for comment_after in line.comments_after(leaf, index):
+        for comment_after in line.comments_after(leaf):
             yield from append_to_line(comment_after)
 
     if current_line:
@@ -2434,6 +2471,16 @@ def is_import(leaf: Leaf) -> bool:
     )
 
 
+def is_special_comment(leaf: Leaf) -> bool:
+    """Return True if the given leaf is a special comment.
+    Only returns true for type comments for now."""
+    t = leaf.type
+    v = leaf.value
+    return bool(
+        (t == token.COMMENT or t == STANDALONE_COMMENT) and (v.startswith("# type:"))
+    )
+
+
 def normalize_prefix(leaf: Leaf, *, inside_brackets: bool) -> None:
     """Leave existing extra newlines if not `inside_brackets`. Remove everything
     else.
@@ -2600,8 +2647,8 @@ def format_int_string(
         return text
 
     text = text.replace("_", "")
-    if len(text) <= 6:
-        # No underscores for numbers <= 6 digits long.
+    if len(text) <= 5:
+        # No underscores for numbers <= 5 digits long.
         return text
 
     if count_from_end:
@@ -2923,6 +2970,7 @@ def ensure_visible(leaf: Leaf) -> None:
 
 def should_explode(line: Line, opening_bracket: Leaf) -> bool:
     """Should `line` immediately be split with `delimiter_split()` after RHS?"""
+
     if not (
         opening_bracket.parent
         and opening_bracket.parent.type in {syms.atom, syms.import_from}
@@ -3405,7 +3453,7 @@ def enumerate_with_length(
             return  # Multiline strings, we can't continue.
 
         comment: Optional[Leaf]
-        for comment in line.comments_after(leaf, index):
+        for comment in line.comments_after(leaf):
             length += len(comment.value)
 
         yield index, leaf, length
@@ -3600,11 +3648,11 @@ def write_cache(
     """Update the cache file."""
     cache_file = get_cache_file(line_length, mode)
     try:
-        if not CACHE_DIR.exists():
-            CACHE_DIR.mkdir(parents=True)
+        CACHE_DIR.mkdir(parents=True, exist_ok=True)
         new_cache = {**cache, **{src.resolve(): get_cache_info(src) for src in sources}}
-        with cache_file.open("wb") as fobj:
-            pickle.dump(new_cache, fobj, protocol=pickle.HIGHEST_PROTOCOL)
+        with tempfile.NamedTemporaryFile(dir=str(cache_file.parent), delete=False) as f:
+            pickle.dump(new_cache, f, protocol=pickle.HIGHEST_PROTOCOL)
+        os.replace(f.name, cache_file)
     except OSError:
         pass
 
@@ -3631,6 +3679,11 @@ def patch_click() -> None:
             module._verify_python3_env = lambda: None
 
 
-if __name__ == "__main__":
+def patched_main() -> None:
+    freeze_support()
     patch_click()
     main()
+
+
+if __name__ == "__main__":
+    patched_main()