]> 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:

Add underscores to numeric literals with more than six digits (#529)
[etc/vim.git] / black.py
index 9b363d5041f2528c063fc70aadc14937f37bb781..56765313aa85b0571c7831f209573bd2b3fbb955 100644 (file)
--- a/black.py
+++ b/black.py
@@ -50,7 +50,7 @@ from blib2to3.pgen2.parse import ParseError
 __version__ = "18.6b4"
 DEFAULT_LINE_LENGTH = 88
 DEFAULT_EXCLUDES = (
-    r"/(\.git|\.hg|\.mypy_cache|\.tox|\.venv|_build|buck-out|build|dist)/"
+    r"/(\.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__))
@@ -79,15 +79,15 @@ syms = pygram.python_symbols
 
 
 class NothingChanged(UserWarning):
-    """Raised by :func:`format_file` when reformatted code is the same as source."""
+    """Raised when reformatted code is the same as source."""
 
 
 class CannotSplit(Exception):
-    """A readable split that fits the allotted line length is impossible.
+    """A readable split that fits the allotted line length is impossible."""
 
-    Raised by :func:`left_hand_split`, :func:`right_hand_split`, and
-    :func:`delimiter_split`.
-    """
+
+class InvalidInput(ValueError):
+    """Raised when input source code fails all parse attempts."""
 
 
 class WriteBack(Enum):
@@ -115,10 +115,16 @@ 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 +133,8 @@ 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
 
 
@@ -170,7 +178,7 @@ def read_pyproject_toml(
     "--line-length",
     type=int,
     default=DEFAULT_LINE_LENGTH,
-    help="How many character per line to allow.",
+    help="How many characters per line to allow.",
     show_default=True,
 )
 @click.option(
@@ -196,6 +204,12 @@ 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 +300,7 @@ def main(
     pyi: bool,
     py36: bool,
     skip_string_normalization: bool,
+    skip_numeric_underscore_normalization: bool,
     quiet: bool,
     verbose: bool,
     include: str,
@@ -296,7 +311,10 @@ 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")
@@ -618,7 +636,8 @@ 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()
@@ -676,7 +695,7 @@ def lib2to3_parse(src_txt: str) -> Node:
                 faulty_line = lines[lineno - 1]
             except IndexError:
                 faulty_line = "<line number missing in source>"
-            exc = ValueError(f"Cannot parse: {lineno}:{column}: {faulty_line}")
+            exc = InvalidInput(f"Cannot parse: {lineno}:{column}: {faulty_line}")
     else:
         raise exc from None
 
@@ -877,8 +896,8 @@ class BracketTracker:
     bracket_match: Dict[Tuple[Depth, NodeType], Leaf] = Factory(dict)
     delimiters: Dict[LeafID, Priority] = Factory(dict)
     previous: Optional[Leaf] = None
-    _for_loop_variable: int = 0
-    _lambda_arguments: int = 0
+    _for_loop_depths: List[int] = Factory(list)
+    _lambda_argument_depths: List[int] = Factory(list)
 
     def mark(self, leaf: Leaf) -> None:
         """Mark `leaf` with bracket-related metadata. Keep track of delimiters.
@@ -951,16 +970,21 @@ class BracketTracker:
         """
         if leaf.type == token.NAME and leaf.value == "for":
             self.depth += 1
-            self._for_loop_variable += 1
+            self._for_loop_depths.append(self.depth)
             return True
 
         return False
 
     def maybe_decrement_after_for_loop_variable(self, leaf: Leaf) -> bool:
         """See `maybe_increment_for_loop_variable` above for explanation."""
-        if self._for_loop_variable and leaf.type == token.NAME and leaf.value == "in":
+        if (
+            self._for_loop_depths
+            and self._for_loop_depths[-1] == self.depth
+            and leaf.type == token.NAME
+            and leaf.value == "in"
+        ):
             self.depth -= 1
-            self._for_loop_variable -= 1
+            self._for_loop_depths.pop()
             return True
 
         return False
@@ -973,16 +997,20 @@ class BracketTracker:
         """
         if leaf.type == token.NAME and leaf.value == "lambda":
             self.depth += 1
-            self._lambda_arguments += 1
+            self._lambda_argument_depths.append(self.depth)
             return True
 
         return False
 
     def maybe_decrement_after_lambda_arguments(self, leaf: Leaf) -> bool:
         """See `maybe_increment_lambda_arguments` above for explanation."""
-        if self._lambda_arguments and leaf.type == token.COLON:
+        if (
+            self._lambda_argument_depths
+            and self._lambda_argument_depths[-1] == self.depth
+            and leaf.type == token.COLON
+        ):
             self.depth -= 1
-            self._lambda_arguments -= 1
+            self._lambda_argument_depths.pop()
             return True
 
         return False
@@ -1164,7 +1192,7 @@ class Line:
             self.remove_trailing_comma()
             return True
 
-        # Otheriwsse, if the trailing one is the only one, we might mistakenly
+        # Otherwise, if the trailing one is the only one, we might mistakenly
         # change a tuple into a different type by removing the comma.
         depth = closing.bracket_depth + 1
         commas = 0
@@ -1377,7 +1405,7 @@ class EmptyLineTracker:
                 newlines = 1
             elif current_line.is_class or self.previous_line.is_class:
                 if current_line.is_stub_class and self.previous_line.is_stub_class:
-                    # No blank line between classes with an emty body
+                    # No blank line between classes with an empty body
                     newlines = 0
                 else:
                     newlines = 1
@@ -1880,7 +1908,7 @@ def is_split_after_delimiter(leaf: Leaf, previous: Leaf = None) -> int:
 
 
 def is_split_before_delimiter(leaf: Leaf, previous: Leaf = None) -> int:
-    """Return the priority of the `leaf` delimiter, given a line before after it.
+    """Return the priority of the `leaf` delimiter, given a line break before it.
 
     The delimiter priorities returned here are from those delimiters that would
     cause a line break before themselves.
@@ -1917,15 +1945,20 @@ def is_split_before_delimiter(leaf: Leaf, previous: Leaf = None) -> int:
     ):
         return STRING_PRIORITY
 
-    if leaf.type != token.NAME:
+    if leaf.type not in {token.NAME, token.ASYNC}:
         return 0
 
     if (
         leaf.value == "for"
         and leaf.parent
         and leaf.parent.type in {syms.comp_for, syms.old_comp_for}
+        or leaf.type == token.ASYNC
     ):
-        return COMPREHENSION_PRIORITY
+        if (
+            not isinstance(leaf.prev_sibling, Leaf)
+            or leaf.prev_sibling.value != "async"
+        ):
+            return COMPREHENSION_PRIORITY
 
     if (
         leaf.value == "if"
@@ -1999,6 +2032,16 @@ def generate_comments(leaf: LN) -> Iterator[Leaf]:
 
 @dataclass
 class ProtoComment:
+    """Describes a piece of syntax that is a comment.
+
+    It's not a :class:`blib2to3.pytree.Leaf` so that:
+
+    * it can be cached (`Leaf` objects should not be reused more than once as
+      they store their lineno, column, prefix, and parent information);
+    * `newlines` and `consumed` fields are kept separate from the `value`. This
+      simplifies handling of special marker comments like ``# fmt: off/on``.
+    """
+
     type: int  # token.COMMENT or STANDALONE_COMMENT
     value: str  # content of the comment
     newlines: int  # how many newlines before the comment
@@ -2007,6 +2050,7 @@ class ProtoComment:
 
 @lru_cache(maxsize=4096)
 def list_comments(prefix: str, *, is_endmarker: bool) -> List[ProtoComment]:
+    """Return a list of :class:`ProtoComment` objects parsed from the given `prefix`."""
     result: List[ProtoComment] = []
     if not prefix or "#" not in prefix:
         return result
@@ -2511,30 +2555,36 @@ def normalize_string_quotes(leaf: Leaf) -> None:
 
 
 def normalize_numeric_literal(leaf: Leaf, allow_underscores: bool) -> None:
-    """Normalizes numeric (float, int, and complex) literals."""
-    # We want all letters (e in exponents, j in complex literals, a-f
-    # in hex literals) to be lowercase.
+    """Normalizes numeric (float, int, and complex) literals.
+
+    All letters used in the representation are normalized to lowercase (except
+    in Python 2 long literals), and long number literals are split using underscores.
+    """
     text = leaf.value.lower()
-    if text.startswith(("0o", "0x", "0b")):
-        # Leave octal, hex, and binary literals alone for now.
+    if text.startswith(("0o", "0b")):
+        # Leave octal and binary literals alone.
         pass
+    elif text.startswith("0x"):
+        # Change hex literals to upper case.
+        before, after = text[:2], text[2:]
+        text = f"{before}{after.upper()}"
     elif "e" in text:
         before, after = text.split("e")
+        sign = ""
         if after.startswith("-"):
             after = after[1:]
             sign = "-"
         elif after.startswith("+"):
             after = after[1:]
-            sign = ""
-        else:
-            sign = ""
         before = format_float_or_int_string(before, allow_underscores)
         after = format_int_string(after, allow_underscores)
         text = f"{before}e{sign}{after}"
-    # Complex numbers and Python 2 longs
-    elif "j" in text or "l" in text:
+    elif text.endswith(("j", "l")):
         number = text[:-1]
         suffix = text[-1]
+        # Capitalize in "2L" because "l" looks too similar to "1".
+        if suffix == "l":
+            suffix = "L"
         text = f"{format_float_or_int_string(number, allow_underscores)}{suffix}"
     else:
         text = format_float_or_int_string(text, allow_underscores)
@@ -2545,24 +2595,40 @@ def format_float_or_int_string(text: str, allow_underscores: bool) -> str:
     """Formats a float string like "1.0"."""
     if "." not in text:
         return format_int_string(text, allow_underscores)
+
     before, after = text.split(".")
     before = format_int_string(before, allow_underscores) if before else "0"
-    after = format_int_string(after, allow_underscores) if after else "0"
+    if after:
+        after = format_int_string(after, allow_underscores, count_from_end=False)
+    else:
+        after = "0"
     return f"{before}.{after}"
 
 
-def format_int_string(text: str, allow_underscores: bool) -> str:
+def format_int_string(
+    text: str, allow_underscores: bool, count_from_end: bool = True
+) -> str:
     """Normalizes underscores in a string to e.g. 1_000_000.
 
-    Input must be a string consisting only of digits and underscores.
+    Input must be a string of digits and optional underscores.
+    If count_from_end is False, we add underscores after groups of three digits
+    counting from the beginning instead of the end of the strings. This is used
+    for the fractional part of float literals.
     """
     if not allow_underscores:
         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
-    return format(int(text), "3_")
+
+    if count_from_end:
+        # Avoid removing leading zeros, which are important if we're formatting
+        # part of a number like "0.001".
+        return format(int("1" + text), "3_")[1:].lstrip("_")
+    else:
+        return "_".join(text[i : i + 3] for i in range(0, len(text), 3))
 
 
 def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
@@ -2583,7 +2649,11 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
     for index, child in enumerate(list(node.children)):
         if check_lpar:
             if child.type == syms.atom:
-                maybe_make_parens_invisible_in_atom(child)
+                if maybe_make_parens_invisible_in_atom(child):
+                    lpar = Leaf(token.LPAR, "")
+                    rpar = Leaf(token.RPAR, "")
+                    index = child.remove() or 0
+                    node.insert_child(index, Node(syms.atom, [lpar, child, rpar]))
             elif is_one_tuple(child):
                 # wrap child in visible parentheses
                 lpar = Leaf(token.LPAR, "(")
@@ -2691,7 +2761,11 @@ def generate_ignored_nodes(leaf: Leaf) -> Iterator[LN]:
 
 
 def maybe_make_parens_invisible_in_atom(node: LN) -> bool:
-    """If it's safe, make the parens in the atom `node` invisible, recursively."""
+    """If it's safe, make the parens in the atom `node` invisible, recursively.
+
+    Returns whether the node should itself be wrapped in invisible parentheses.
+
+    """
     if (
         node.type != syms.atom
         or is_empty_tuple(node)
@@ -2709,9 +2783,9 @@ def maybe_make_parens_invisible_in_atom(node: LN) -> bool:
         last.value = ""  # type: ignore
         if len(node.children) > 1:
             maybe_make_parens_invisible_in_atom(node.children[1])
-        return True
+        return False
 
-    return False
+    return True
 
 
 def is_empty_tuple(node: LN) -> bool:
@@ -2889,7 +2963,8 @@ def is_python36(node: Node) -> bool:
     """Return True if the current file is using Python 3.6+ features.
 
     Currently looking for:
-    - f-strings; and
+    - f-strings;
+    - underscores in numeric literals; and
     - trailing commas after * or ** in function signatures and calls.
     """
     for n in node.pre_order():
@@ -2898,6 +2973,10 @@ def is_python36(node: Node) -> bool:
             if value_head in {'f"', 'F"', "f'", "F'", "rf", "fr", "RF", "FR"}:
                 return True
 
+        elif n.type == token.NUMBER:
+            if "_" in n.value:  # type: ignore
+                return True
+
         elif (
             n.type in {syms.typedargslist, syms.arglist}
             and n.children
@@ -2931,7 +3010,6 @@ def generate_trailers_to_omit(line: Line, line_length: int) -> Iterator[Set[Leaf
     length = 4 * line.depth
     opening_bracket = None
     closing_bracket = None
-    optional_brackets: Set[LeafID] = set()
     inner_brackets: Set[LeafID] = set()
     for index, leaf, leaf_length in enumerate_with_length(line, reversed=True):
         length += leaf_length
@@ -2942,17 +3020,12 @@ def generate_trailers_to_omit(line: Line, line_length: int) -> Iterator[Set[Leaf
         if leaf.type == STANDALONE_COMMENT or has_inline_comment:
             break
 
-        optional_brackets.discard(id(leaf))
         if opening_bracket:
             if leaf is opening_bracket:
                 opening_bracket = None
             elif leaf.type in CLOSING_BRACKETS:
                 inner_brackets.add(id(leaf))
         elif leaf.type in CLOSING_BRACKETS:
-            if not leaf.value:
-                optional_brackets.add(id(opening_bracket))
-                continue
-
             if index > 0 and line.leaves[index - 1].type in OPENING_BRACKETS:
                 # Empty brackets would fail a split so treat them as "inner"
                 # brackets (e.g. only add them to the `omit` set if another
@@ -2960,13 +3033,15 @@ def generate_trailers_to_omit(line: Line, line_length: int) -> Iterator[Set[Leaf
                 inner_brackets.add(id(leaf))
                 continue
 
-            opening_bracket = leaf.opening_bracket
             if closing_bracket:
                 omit.add(id(closing_bracket))
                 omit.update(inner_brackets)
                 inner_brackets.clear()
                 yield omit
-            closing_bracket = leaf
+
+            if leaf.value:
+                opening_bracket = leaf.opening_bracket
+                closing_bracket = leaf
 
 
 def get_future_imports(node: Node) -> Set[str]:
@@ -3131,7 +3206,7 @@ class Report:
         - otherwise return 0.
         """
         # According to http://tldp.org/LDP/abs/html/exitcodes.html starting with
-        # 126 we have special returncodes reserved by the shell.
+        # 126 we have special return codes reserved by the shell.
         if self.failure_count:
             return 123