]> git.madduck.net Git - etc/vim.git/blobdiff - src/black/linegen.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:

Remove unnecessary parentheses from `except` clauses (#2939)
[etc/vim.git] / src / black / linegen.py
index eb53fa0ac56e0aa2520d3b27248d3e4123c7c7ef..8a28c3901bbeb90dd68806f4e0adaac7241ab861 100644 (file)
@@ -5,12 +5,16 @@ from functools import partial, wraps
 import sys
 from typing import Collection, Iterator, List, Optional, Set, Union
 
-from dataclasses import dataclass, field
-
 from black.nodes import WHITESPACE, RARROW, STATEMENT, STANDALONE_COMMENT
 from black.nodes import ASSIGNMENTS, OPENING_BRACKETS, CLOSING_BRACKETS
-from black.nodes import Visitor, syms, first_child_is_arith, ensure_visible
-from black.nodes import is_docstring, is_empty_tuple, is_one_tuple, is_one_tuple_between
+from black.nodes import Visitor, syms, is_arith_like, ensure_visible
+from black.nodes import (
+    is_docstring,
+    is_empty_tuple,
+    is_one_tuple,
+    is_one_sequence_between,
+)
+from black.nodes import is_name_token, is_lpar_token, is_rpar_token
 from black.nodes import is_walrus_assignment, is_yield, is_vararg, is_multiline_string
 from black.nodes import is_stub_suite, is_stub_body, is_atom_with_invisible_parens
 from black.nodes import wrap_in_parentheses
@@ -22,10 +26,9 @@ from black.comments import generate_comments, list_comments, FMT_OFF
 from black.numerics import normalize_numeric_literal
 from black.strings import get_string_prefix, fix_docstring
 from black.strings import normalize_string_prefix, normalize_string_quotes
-from black.trans import Transformer, CannotTransform, StringMerger
-from black.trans import StringSplitter, StringParenWrapper, StringParenStripper
-from black.mode import Mode
-from black.mode import Feature
+from black.trans import Transformer, CannotTransform, StringMerger, StringSplitter
+from black.trans import StringParenWrapper, StringParenStripper, hug_power_op
+from black.mode import Mode, Feature, Preview
 
 from blib2to3.pytree import Node, Leaf
 from blib2to3.pgen2 import token
@@ -40,7 +43,8 @@ class CannotSplit(CannotTransform):
     """A readable split that fits the allotted line length is impossible."""
 
 
-@dataclass
+# This isn't a dataclass because @dataclass + Generic breaks mypyc.
+# See also https://github.com/mypyc/mypyc/issues/827.
 class LineGenerator(Visitor[Line]):
     """Generates reformatted Line objects.  Empty lines are not emitted.
 
@@ -48,9 +52,10 @@ class LineGenerator(Visitor[Line]):
     in ways that will no longer stringify to valid Python code on the tree.
     """
 
-    mode: Mode
-    remove_u_prefix: bool = False
-    current_line: Line = field(init=False)
+    def __init__(self, mode: Mode) -> None:
+        self.mode = mode
+        self.current_line: Line
+        self.__post_init__()
 
     def line(self, indent: int = 0) -> Iterator[Line]:
         """Generate a line.
@@ -72,7 +77,7 @@ class LineGenerator(Visitor[Line]):
         """Default `visit_*()` implementation. Recurses to children of `node`."""
         if isinstance(node, Leaf):
             any_open_brackets = self.current_line.bracket_tracker.any_open_brackets()
-            for comment in generate_comments(node):
+            for comment in generate_comments(node, preview=self.mode.preview):
                 if any_open_brackets:
                     # any comment within brackets is subject to splitting
                     self.current_line.append(comment)
@@ -90,9 +95,7 @@ class LineGenerator(Visitor[Line]):
 
             normalize_prefix(node, inside_brackets=any_open_brackets)
             if self.mode.string_normalization and node.type == token.STRING:
-                node.value = normalize_string_prefix(
-                    node.value, remove_u_prefix=self.remove_u_prefix
-                )
+                node.value = normalize_string_prefix(node.value)
                 node.value = normalize_string_quotes(node.value)
             if node.type == token.NUMBER:
                 normalize_numeric_literal(node)
@@ -126,7 +129,7 @@ class LineGenerator(Visitor[Line]):
         """Visit a statement.
 
         This implementation is shared for `if`, `while`, `for`, `try`, `except`,
-        `def`, `with`, `class`, `assert` and assignments.
+        `def`, `with`, `class`, `assert`, and assignments.
 
         The relevant Python language `keywords` for a given statement will be
         NAME leaves within it. This methods puts those on a separate line.
@@ -134,13 +137,21 @@ class LineGenerator(Visitor[Line]):
         `parens` holds a set of string leaf values immediately after which
         invisible parens should be put.
         """
-        normalize_invisible_parens(node, parens_after=parens)
+        normalize_invisible_parens(node, parens_after=parens, preview=self.mode.preview)
         for child in node.children:
-            if child.type == token.NAME and child.value in keywords:  # type: ignore
+            if is_name_token(child) and child.value in keywords:
                 yield from self.line()
 
             yield from self.visit(child)
 
+    def visit_match_case(self, node: Node) -> Iterator[Line]:
+        """Visit either a match or case statement."""
+        normalize_invisible_parens(node, parens_after=set(), preview=self.mode.preview)
+
+        yield from self.line()
+        for child in node.children:
+            yield from self.visit(child)
+
     def visit_suite(self, node: Node) -> Iterator[Line]:
         """Visit a suite."""
         if self.mode.is_pyi and is_stub_suite(node):
@@ -150,8 +161,12 @@ class LineGenerator(Visitor[Line]):
 
     def visit_simple_stmt(self, node: Node) -> Iterator[Line]:
         """Visit a statement without nested statements."""
-        if first_child_is_arith(node):
-            wrap_in_parentheses(node, node.children[0], visible=False)
+        prev_type: Optional[int] = None
+        for child in node.children:
+            if (prev_type is None or prev_type == token.SEMI) and is_arith_like(child):
+                wrap_in_parentheses(node, child, visible=False)
+            prev_type = child.type
+
         is_suite_like = node.parent and node.parent.type in STATEMENT
         if is_suite_like:
             if self.mode.is_pyi and is_stub_body(node):
@@ -191,6 +206,28 @@ class LineGenerator(Visitor[Line]):
             yield from self.line()
             yield from self.visit(child)
 
+    def visit_power(self, node: Node) -> Iterator[Line]:
+        for idx, leaf in enumerate(node.children[:-1]):
+            next_leaf = node.children[idx + 1]
+
+            if not isinstance(leaf, Leaf):
+                continue
+
+            value = leaf.value.lower()
+            if (
+                leaf.type == token.NUMBER
+                and next_leaf.type == syms.trailer
+                # Ensure that we are in an attribute trailer
+                and next_leaf.children[0].type == token.DOT
+                # It shouldn't wrap hexadecimal, binary and octal literals
+                and not value.startswith(("0x", "0b", "0o"))
+                # It shouldn't wrap complex literals
+                and "j" not in value
+            ):
+                wrap_in_parentheses(node, leaf)
+
+        yield from self.visit_default(node)
+
     def visit_SEMI(self, leaf: Leaf) -> Iterator[Line]:
         """Remove a semicolon and put the other statement on a separate line."""
         yield from self.line()
@@ -226,7 +263,7 @@ class LineGenerator(Visitor[Line]):
         if is_docstring(leaf) and "\\\n" not in leaf.value:
             # We're ignoring docstrings with backslash newline escapes because changing
             # indentation of those changes the AST representation of the code.
-            docstring = normalize_string_prefix(leaf.value, self.remove_u_prefix)
+            docstring = normalize_string_prefix(leaf.value)
             prefix = get_string_prefix(docstring)
             docstring = docstring[len(prefix) :]  # Remove the prefix
             quote_char = docstring[0]
@@ -281,7 +318,12 @@ class LineGenerator(Visitor[Line]):
         self.visit_try_stmt = partial(
             v, keywords={"try", "except", "else", "finally"}, parens=Ø
         )
-        self.visit_except_clause = partial(v, keywords={"except"}, parens=Ø)
+        if self.mode.preview:
+            self.visit_except_clause = partial(
+                v, keywords={"except"}, parens={"except"}
+            )
+        else:
+            self.visit_except_clause = partial(v, keywords={"except"}, parens=Ø)
         self.visit_with_stmt = partial(v, keywords={"with"}, parens=Ø)
         self.visit_funcdef = partial(v, keywords={"def"}, parens=Ø)
         self.visit_classdef = partial(v, keywords={"class"}, parens=Ø)
@@ -292,6 +334,10 @@ class LineGenerator(Visitor[Line]):
         self.visit_async_funcdef = self.visit_async_stmt
         self.visit_decorated = self.visit_decorators
 
+        # PEP 634
+        self.visit_match_stmt = self.visit_match_case
+        self.visit_case_block = self.visit_match_case
+
 
 def transform_line(
     line: Line, mode: Mode, features: Collection[Feature] = ()
@@ -327,7 +373,7 @@ def transform_line(
         and not (line.inside_brackets and line.contains_standalone_comments())
     ):
         # Only apply basic string preprocessing, since lines shouldn't be split here.
-        if mode.experimental_string_processing:
+        if Preview.string_processing in mode:
             transformers = [string_merge, string_paren_strip]
         else:
             transformers = []
@@ -335,7 +381,9 @@ def transform_line(
         transformers = [left_hand_split]
     else:
 
-        def rhs(line: Line, features: Collection[Feature]) -> Iterator[Line]:
+        def _rhs(
+            self: object, line: Line, features: Collection[Feature]
+        ) -> Iterator[Line]:
             """Wraps calls to `right_hand_split`.
 
             The calls increasingly `omit` right-hand trailers (bracket pairs with
@@ -362,7 +410,13 @@ def transform_line(
                 line, line_length=mode.line_length, features=features
             )
 
-        if mode.experimental_string_processing:
+        # HACK: nested functions (like _rhs) compiled by mypyc don't retain their
+        # __name__ attribute which is needed in `run_transformer` further down.
+        # Unfortunately a nested class breaks mypyc too. So a class must be created
+        # via type ... https://github.com/mypyc/mypyc/issues/884
+        rhs = type("rhs", (), {"__call__": _rhs})()
+
+        if Preview.string_processing in mode:
             if line.inside_brackets:
                 transformers = [
                     string_merge,
@@ -386,6 +440,9 @@ def transform_line(
                 transformers = [delimiter_split, standalone_comment_split, rhs]
             else:
                 transformers = [rhs]
+    # It's always safe to attempt hugging of power operations and pretty much every line
+    # could match.
+    transformers.append(hug_power_op)
 
     for transform in transformers:
         # We are accumulating lines in `result` because we might want to abort
@@ -496,7 +553,7 @@ def right_hand_split(
         # 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_on_explode=omit)
+        and can_omit_invisible_parens(body, line_length)
     ):
         omit = {id(closing_bracket), *omit}
         try:
@@ -618,9 +675,9 @@ def dont_increase_indentation(split_func: Transformer) -> Transformer:
 
     @wraps(split_func)
     def split_wrapper(line: Line, features: Collection[Feature] = ()) -> Iterator[Line]:
-        for line in split_func(line, features):
-            normalize_prefix(line.leaves[0], inside_brackets=True)
-            yield line
+        for split_line in split_func(line, features):
+            normalize_prefix(split_line.leaves[0], inside_brackets=True)
+            yield split_line
 
     return split_wrapper
 
@@ -755,7 +812,9 @@ def normalize_prefix(leaf: Leaf, *, inside_brackets: bool) -> None:
     leaf.prefix = ""
 
 
-def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
+def normalize_invisible_parens(
+    node: Node, parens_after: Set[str], *, preview: bool
+) -> None:
     """Make existing optional parentheses invisible or create new ones.
 
     `parens_after` is a set of string leaf values immediately after which parens
@@ -764,7 +823,7 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
     Standardizes on visible parentheses for single-element tuples, and keeps
     existing visible parentheses for other tuples and generator expressions.
     """
-    for pc in list_comments(node.prefix, is_endmarker=False):
+    for pc in list_comments(node.prefix, is_endmarker=False, preview=preview):
         if pc.value in FMT_OFF:
             # This `node` has a prefix with `# fmt: off`, don't mess with parens.
             return
@@ -773,7 +832,9 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
         # Fixes a bug where invisible parens are not properly stripped from
         # assignment statements that contain type annotations.
         if isinstance(child, Node) and child.type == syms.annassign:
-            normalize_invisible_parens(child, parens_after=parens_after)
+            normalize_invisible_parens(
+                child, parens_after=parens_after, preview=preview
+            )
 
         # Add parentheses around long tuple unpacking in assignments.
         if (
@@ -785,17 +846,22 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
 
         if check_lpar:
             if child.type == syms.atom:
-                if maybe_make_parens_invisible_in_atom(child, parent=node):
+                if maybe_make_parens_invisible_in_atom(
+                    child,
+                    parent=node,
+                    preview=preview,
+                ):
                     wrap_in_parentheses(node, child, visible=False)
             elif is_one_tuple(child):
                 wrap_in_parentheses(node, child, visible=True)
             elif node.type == syms.import_from:
                 # "import from" nodes store parentheses directly as part of
                 # the statement
-                if child.type == token.LPAR:
+                if is_lpar_token(child):
+                    assert is_rpar_token(node.children[-1])
                     # make parentheses invisible
-                    child.value = ""  # type: ignore
-                    node.children[-1].value = ""  # type: ignore
+                    child.value = ""
+                    node.children[-1].value = ""
                 elif child.type != token.STAR:
                     # insert invisible parentheses
                     node.insert_child(index, Leaf(token.LPAR, ""))
@@ -808,7 +874,11 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
         check_lpar = isinstance(child, Leaf) and child.value in parens_after
 
 
-def maybe_make_parens_invisible_in_atom(node: LN, parent: LN) -> bool:
+def maybe_make_parens_invisible_in_atom(
+    node: LN,
+    parent: LN,
+    preview: bool = False,
+) -> bool:
     """If it's safe, make the parens in the atom `node` invisible, recursively.
     Additionally, remove repeated, adjacent invisible parens from the atom `node`
     as they are redundant.
@@ -816,13 +886,23 @@ def maybe_make_parens_invisible_in_atom(node: LN, parent: LN) -> bool:
     Returns whether the node should itself be wrapped in invisible parentheses.
 
     """
+    if (
+        preview
+        and parent.type == syms.for_stmt
+        and isinstance(node.prev_sibling, Leaf)
+        and node.prev_sibling.type == token.NAME
+        and node.prev_sibling.value == "for"
+    ):
+        for_stmt_check = False
+    else:
+        for_stmt_check = True
 
     if (
         node.type != syms.atom
         or is_empty_tuple(node)
         or is_one_tuple(node)
         or (is_yield(node) and parent.type != syms.expr_stmt)
-        or max_delimiter_priority_in_atom(node) >= COMMA_PRIORITY
+        or (max_delimiter_priority_in_atom(node) >= COMMA_PRIORITY and for_stmt_check)
     ):
         return False
 
@@ -840,12 +920,12 @@ def maybe_make_parens_invisible_in_atom(node: LN, parent: LN) -> bool:
 
     first = node.children[0]
     last = node.children[-1]
-    if first.type == token.LPAR and last.type == token.RPAR:
+    if is_lpar_token(first) and is_rpar_token(last):
         middle = node.children[1]
         # make parentheses invisible
-        first.value = ""  # type: ignore
-        last.value = ""  # type: ignore
-        maybe_make_parens_invisible_in_atom(middle, parent=parent)
+        first.value = ""
+        last.value = ""
+        maybe_make_parens_invisible_in_atom(middle, parent=parent, preview=preview)
 
         if is_atom_with_invisible_parens(middle):
             # Strip the invisible parens from `middle` by replacing
@@ -920,7 +1000,8 @@ def generate_trailers_to_omit(line: Line, line_length: int) -> Iterator[Set[Leaf
                 if (
                     prev
                     and prev.type == token.COMMA
-                    and not is_one_tuple_between(
+                    and leaf.opening_bracket is not None
+                    and not is_one_sequence_between(
                         leaf.opening_bracket, leaf, line.leaves
                     )
                 ):
@@ -947,7 +1028,8 @@ def generate_trailers_to_omit(line: Line, line_length: int) -> Iterator[Set[Leaf
             if (
                 prev
                 and prev.type == token.COMMA
-                and not is_one_tuple_between(leaf.opening_bracket, leaf, line.leaves)
+                and leaf.opening_bracket is not None
+                and not is_one_sequence_between(leaf.opening_bracket, leaf, line.leaves)
             ):
                 # Never omit bracket pairs with trailing commas.
                 # We need to explode on those.
@@ -976,7 +1058,7 @@ def run_transformer(
         result.extend(transform_line(transformed_line, mode=mode, features=features))
 
     if (
-        transform.__name__ != "rhs"
+        transform.__class__.__name__ != "rhs"
         or not line.bracket_tracker.invisible
         or any(bracket.value for bracket in line.bracket_tracker.invisible)
         or line.contains_multiline_strings()