]> 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:

Move test for g:load_black to improve plugin performance (GH-2896)
[etc/vim.git] / src / black / linegen.py
index 8cf32c973bb6a4880e543ec6a51ec06ff54dc492..4dc242a1dfe05b9f8ef7ff53b26333073469cf57 100644 (file)
@@ -5,12 +5,11 @@ from functools import partial, wraps
 import sys
 from typing import Collection, Iterator, List, Optional, Set, Union
 
 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 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 Visitor, syms, is_arith_like, ensure_visible
 from black.nodes import is_docstring, is_empty_tuple, is_one_tuple, is_one_tuple_between
 from black.nodes import is_docstring, is_empty_tuple, is_one_tuple, is_one_tuple_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
 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 +21,9 @@ 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.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
 
 from blib2to3.pytree import Node, Leaf
 from blib2to3.pgen2 import token
@@ -40,7 +38,8 @@ class CannotSplit(CannotTransform):
     """A readable split that fits the allotted line length is impossible."""
 
 
     """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.
 
 class LineGenerator(Visitor[Line]):
     """Generates reformatted Line objects.  Empty lines are not emitted.
 
@@ -48,9 +47,10 @@ class LineGenerator(Visitor[Line]):
     in ways that will no longer stringify to valid Python code on the tree.
     """
 
     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.
 
     def line(self, indent: int = 0) -> Iterator[Line]:
         """Generate a line.
@@ -90,9 +90,7 @@ class LineGenerator(Visitor[Line]):
 
             normalize_prefix(node, inside_brackets=any_open_brackets)
             if self.mode.string_normalization and node.type == token.STRING:
 
             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)
                 node.value = normalize_string_quotes(node.value)
             if node.type == token.NUMBER:
                 normalize_numeric_literal(node)
@@ -126,7 +124,7 @@ class LineGenerator(Visitor[Line]):
         """Visit a statement.
 
         This implementation is shared for `if`, `while`, `for`, `try`, `except`,
         """Visit a statement.
 
         This implementation is shared for `if`, `while`, `for`, `try`, `except`,
-        `def`, `with`, `class`, `assert`, `match`, `case` 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.
 
         The relevant Python language `keywords` for a given statement will be
         NAME leaves within it. This methods puts those on a separate line.
@@ -136,11 +134,19 @@ class LineGenerator(Visitor[Line]):
         """
         normalize_invisible_parens(node, parens_after=parens)
         for child in node.children:
         """
         normalize_invisible_parens(node, parens_after=parens)
         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)
 
                 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())
+
+        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):
     def visit_suite(self, node: Node) -> Iterator[Line]:
         """Visit a suite."""
         if self.mode.is_pyi and is_stub_suite(node):
@@ -150,8 +156,12 @@ class LineGenerator(Visitor[Line]):
 
     def visit_simple_stmt(self, node: Node) -> Iterator[Line]:
         """Visit a statement without nested statements."""
 
     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):
         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 +201,28 @@ class LineGenerator(Visitor[Line]):
             yield from self.line()
             yield from self.visit(child)
 
             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()
     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 +258,7 @@ 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.
         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]
             prefix = get_string_prefix(docstring)
             docstring = docstring[len(prefix) :]  # Remove the prefix
             quote_char = docstring[0]
@@ -293,8 +325,8 @@ class LineGenerator(Visitor[Line]):
         self.visit_decorated = self.visit_decorators
 
         # PEP 634
         self.visit_decorated = self.visit_decorators
 
         # PEP 634
-        self.visit_match_stmt = partial(v, keywords={"match"}, parens=Ø)
-        self.visit_case_block = partial(v, keywords={"case"}, parens=Ø)
+        self.visit_match_stmt = self.visit_match_case
+        self.visit_case_block = self.visit_match_case
 
 
 def transform_line(
 
 
 def transform_line(
@@ -331,7 +363,7 @@ 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.
         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 = []
             transformers = [string_merge, string_paren_strip]
         else:
             transformers = []
@@ -339,7 +371,9 @@ def transform_line(
         transformers = [left_hand_split]
     else:
 
         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
             """Wraps calls to `right_hand_split`.
 
             The calls increasingly `omit` right-hand trailers (bracket pairs with
@@ -366,7 +400,13 @@ def transform_line(
                 line, line_length=mode.line_length, features=features
             )
 
                 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,
             if line.inside_brackets:
                 transformers = [
                     string_merge,
@@ -390,6 +430,9 @@ def transform_line(
                 transformers = [delimiter_split, standalone_comment_split, rhs]
             else:
                 transformers = [rhs]
                 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
 
     for transform in transformers:
         # We are accumulating lines in `result` because we might want to abort
@@ -500,7 +543,7 @@ 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
         # 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:
     ):
         omit = {id(closing_bracket), *omit}
         try:
@@ -796,10 +839,11 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
             elif node.type == syms.import_from:
                 # "import from" nodes store parentheses directly as part of
                 # the statement
             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
                     # 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, ""))
                 elif child.type != token.STAR:
                     # insert invisible parentheses
                     node.insert_child(index, Leaf(token.LPAR, ""))
@@ -844,11 +888,11 @@ def maybe_make_parens_invisible_in_atom(node: LN, parent: LN) -> bool:
 
     first = node.children[0]
     last = node.children[-1]
 
     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
         middle = node.children[1]
         # make parentheses invisible
-        first.value = ""  # type: ignore
-        last.value = ""  # type: ignore
+        first.value = ""
+        last.value = ""
         maybe_make_parens_invisible_in_atom(middle, parent=parent)
 
         if is_atom_with_invisible_parens(middle):
         maybe_make_parens_invisible_in_atom(middle, parent=parent)
 
         if is_atom_with_invisible_parens(middle):
@@ -924,6 +968,7 @@ def generate_trailers_to_omit(line: Line, line_length: int) -> Iterator[Set[Leaf
                 if (
                     prev
                     and prev.type == token.COMMA
                 if (
                     prev
                     and prev.type == token.COMMA
+                    and leaf.opening_bracket is not None
                     and not is_one_tuple_between(
                         leaf.opening_bracket, leaf, line.leaves
                     )
                     and not is_one_tuple_between(
                         leaf.opening_bracket, leaf, line.leaves
                     )
@@ -951,6 +996,7 @@ def generate_trailers_to_omit(line: Line, line_length: int) -> Iterator[Set[Leaf
             if (
                 prev
                 and prev.type == token.COMMA
             if (
                 prev
                 and prev.type == token.COMMA
+                and leaf.opening_bracket is not None
                 and not is_one_tuple_between(leaf.opening_bracket, leaf, line.leaves)
             ):
                 # Never omit bracket pairs with trailing commas.
                 and not is_one_tuple_between(leaf.opening_bracket, leaf, line.leaves)
             ):
                 # Never omit bracket pairs with trailing commas.
@@ -980,7 +1026,7 @@ def run_transformer(
         result.extend(transform_line(transformed_line, mode=mode, features=features))
 
     if (
         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()
         or not line.bracket_tracker.invisible
         or any(bracket.value for bracket in line.bracket_tracker.invisible)
         or line.contains_multiline_strings()