Don't let TokenError bubble up from lib2to3_parse (GH-2343)

[etc/vim.git] / src / black / linegen.py

diff --git a/src/black/linegen.py b/src/black/linegen.py
index 2e16b6fde4457d24c06992a57d1324e63675d9e4..f234913a161452e69747666b696d9e299ed0422d 100644 (file)
--- a/src/black/linegen.py
+++ b/src/black/linegen.py
@@ -5,9 +5,7 @@ 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, STATEMENT, STANDALONE_COMMENT
+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 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


@@ -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 +47,11 @@ 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, remove_u_prefix: bool = False) -> None:
+        self.mode = mode
+        self.remove_u_prefix = remove_u_prefix
+        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.


@@ -126,7 +127,7 @@ 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` 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.


@@ -141,6 +142,14 @@ class LineGenerator(Visitor[Line]):
 
             yield from self.visit(child)
 
 
             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):


@@ -226,8 +235,9 @@ 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.

-            prefix = get_string_prefix(leaf.value)
-            docstring = leaf.value[len(prefix) :]  # Remove the prefix
+            docstring = normalize_string_prefix(leaf.value, self.remove_u_prefix)
+            prefix = get_string_prefix(docstring)
+            docstring = docstring[len(prefix) :]  # Remove the prefix

             quote_char = docstring[0]
             # A natural way to remove the outer quotes is to do:
             #   docstring = docstring.strip(quote_char)
             quote_char = docstring[0]
             # A natural way to remove the outer quotes is to do:
             #   docstring = docstring.strip(quote_char)


@@ -236,6 +246,7 @@ class LineGenerator(Visitor[Line]):
             # characters but only if they are the same as the first.
             quote_len = 1 if docstring[1] != quote_char else 3
             docstring = docstring[quote_len:-quote_len]
             # characters but only if they are the same as the first.
             quote_len = 1 if docstring[1] != quote_char else 3
             docstring = docstring[quote_len:-quote_len]

+            docstring_started_empty = not docstring

 
             if is_multiline_string(leaf):
                 indent = " " * 4 * self.current_line.depth
 
             if is_multiline_string(leaf):
                 indent = " " * 4 * self.current_line.depth


@@ -255,8 +266,7 @@ class LineGenerator(Visitor[Line]):
                         # Odd number of tailing backslashes, add some padding to
                         # avoid escaping the closing string quote.
                         docstring += " "
                         # Odd number of tailing backslashes, add some padding to
                         # avoid escaping the closing string quote.
                         docstring += " "

-            else:
-                # Add some padding if the docstring is empty.
+            elif not docstring_started_empty:

                 docstring = " "
 
             # We could enforce triple quotes at this point.
                 docstring = " "
 
             # We could enforce triple quotes at this point.


@@ -291,6 +301,10 @@ class LineGenerator(Visitor[Line]):
         self.visit_async_funcdef = self.visit_async_stmt
         self.visit_decorated = self.visit_decorators
 
         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] = ()
 
 def transform_line(
     line: Line, mode: Mode, features: Collection[Feature] = ()


@@ -334,7 +348,9 @@ 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


@@ -361,6 +377,12 @@ def transform_line(
                 line, line_length=mode.line_length, features=features
             )
 
                 line, line_length=mode.line_length, features=features
             )
 

+        # 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 mode.experimental_string_processing:
             if line.inside_brackets:
                 transformers = [
         if mode.experimental_string_processing:
             if line.inside_brackets:
                 transformers = [


@@ -502,14 +524,14 @@ def right_hand_split(
             yield from right_hand_split(line, line_length, features=features, omit=omit)
             return
 
             yield from right_hand_split(line, line_length, features=features, omit=omit)
             return
 

-        except CannotSplit:
+        except CannotSplit as e:

             if not (
                 can_be_split(body)
                 or is_line_short_enough(body, line_length=line_length)
             ):
                 raise CannotSplit(
                     "Splitting failed, body is still too long and can't be split."
             if not (
                 can_be_split(body)
                 or is_line_short_enough(body, line_length=line_length)
             ):
                 raise CannotSplit(
                     "Splitting failed, body is still too long and can't be split."

-                )
+                ) from e

 
             elif head.contains_multiline_strings() or tail.contains_multiline_strings():
                 raise CannotSplit(
 
             elif head.contains_multiline_strings() or tail.contains_multiline_strings():
                 raise CannotSplit(


@@ -517,7 +539,7 @@ def right_hand_split(
                     " satisfy the splitting algorithm because the head or the tail"
                     " contains multiline strings which by definition never fit one"
                     " line."
                     " satisfy the splitting algorithm because the head or the tail"
                     " contains multiline strings which by definition never fit one"
                     " line."

-                )
+                ) from e

 
     ensure_visible(opening_bracket)
     ensure_visible(closing_bracket)
 
     ensure_visible(opening_bracket)
     ensure_visible(closing_bracket)


@@ -573,6 +595,20 @@ def bracket_split_build_line(
                 original.is_def
                 and opening_bracket.value == "("
                 and not any(leaf.type == token.COMMA for leaf in leaves)
                 original.is_def
                 and opening_bracket.value == "("
                 and not any(leaf.type == token.COMMA for leaf in leaves)

+                # In particular, don't add one within a parenthesized return annotation.
+                # Unfortunately the indicator we're in a return annotation (RARROW) may
+                # be defined directly in the parent node, the parent of the parent ...
+                # and so on depending on how complex the return annotation is.
+                # This isn't perfect and there's some false negatives but they are in
+                # contexts were a comma is actually fine.
+                and not any(
+                    node.prev_sibling.type == RARROW
+                    for node in (
+                        leaves[0].parent,
+                        getattr(leaves[0].parent, "parent", None),
+                    )
+                    if isinstance(node, Node) and isinstance(node.prev_sibling, Leaf)
+                )

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


@@ -620,13 +656,13 @@ def delimiter_split(line: Line, features: Collection[Feature] = ()) -> Iterator[
     try:
         last_leaf = line.leaves[-1]
     except IndexError:
     try:
         last_leaf = line.leaves[-1]
     except IndexError:

-        raise CannotSplit("Line empty")
+        raise CannotSplit("Line empty") from None

 
     bt = line.bracket_tracker
     try:
         delimiter_priority = bt.max_delimiter_priority(exclude={id(last_leaf)})
     except ValueError:
 
     bt = line.bracket_tracker
     try:
         delimiter_priority = bt.max_delimiter_priority(exclude={id(last_leaf)})
     except ValueError:

-        raise CannotSplit("No delimiters found")
+        raise CannotSplit("No delimiters found") from None

 
     if delimiter_priority == DOT_PRIORITY:
         if bt.delimiter_count_with_priority(delimiter_priority) == 1:
 
     if delimiter_priority == DOT_PRIORITY:
         if bt.delimiter_count_with_priority(delimiter_priority) == 1:


@@ -960,14 +996,19 @@ def run_transformer(
 
         result.extend(transform_line(transformed_line, mode=mode, features=features))
 
 
         result.extend(transform_line(transformed_line, mode=mode, features=features))
 

-    if not (
-        transform.__name__ == "rhs"
-        and line.bracket_tracker.invisible
-        and not any(bracket.value for bracket in line.bracket_tracker.invisible)
-        and not line.contains_multiline_strings()
-        and not result[0].contains_uncollapsable_type_comments()
-        and not result[0].contains_unsplittable_type_ignore()
-        and not is_line_short_enough(result[0], line_length=mode.line_length)
+    if (
+        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 result[0].contains_uncollapsable_type_comments()
+        or result[0].contains_unsplittable_type_ignore()
+        or is_line_short_enough(result[0], line_length=mode.line_length)
+        # If any leaves have no parents (which _can_ occur since
+        # `transform(line)` potentially destroys the line's underlying node
+        # structure), then we can't proceed. Doing so would cause the below
+        # call to `append_leaves()` to fail.
+        or any(leaf.parent is None for leaf in line.leaves)

     ):
         return result
 
     ):
         return result