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

Fix a crash in ESP where a standalone comment is placed before a dict's value (#3469)
[etc/vim.git] / src / black / trans.py
index 01aa80eaaf84760f765ae34eef651d7336e28c09..0eb53e2b09877652bcfa12e6399dbd1937930aa1 100644 (file)
@@ -1,10 +1,11 @@
 """
 String transformers that can split and merge strings.
 """
+import re
+import sys
 from abc import ABC, abstractmethod
 from collections import defaultdict
 from dataclasses import dataclass
-import re
 from typing import (
     Any,
     Callable,
@@ -21,29 +22,38 @@ from typing import (
     TypeVar,
     Union,
 )
-import sys
 
 if sys.version_info < (3, 8):
-    from typing_extensions import Literal, Final
+    from typing_extensions import Final, Literal
 else:
     from typing import Literal, Final
 
 from mypy_extensions import trait
 
-from black.rusty import Result, Ok, Err
-
-from black.mode import Feature
-from black.nodes import syms, replace_child, parent_type
-from black.nodes import is_empty_par, is_empty_lpar, is_empty_rpar
-from black.nodes import OPENING_BRACKETS, CLOSING_BRACKETS, STANDALONE_COMMENT
-from black.lines import Line, append_leaves
-from black.brackets import BracketMatchError
 from black.comments import contains_pragma_comment
-from black.strings import has_triple_quotes, get_string_prefix, assert_is_leaf_string
-from black.strings import normalize_string_quotes
-
-from blib2to3.pytree import Leaf, Node
+from black.lines import Line, append_leaves
+from black.mode import Feature
+from black.nodes import (
+    CLOSING_BRACKETS,
+    OPENING_BRACKETS,
+    STANDALONE_COMMENT,
+    is_empty_lpar,
+    is_empty_par,
+    is_empty_rpar,
+    is_part_of_annotation,
+    parent_type,
+    replace_child,
+    syms,
+)
+from black.rusty import Err, Ok, Result
+from black.strings import (
+    assert_is_leaf_string,
+    get_string_prefix,
+    has_triple_quotes,
+    normalize_string_quotes,
+)
 from blib2to3.pgen2 import token
+from blib2to3.pytree import Leaf, Node
 
 
 class CannotTransform(Exception):
@@ -121,7 +131,7 @@ def hug_power_op(line: Line, features: Collection[Feature]) -> Iterator[Line]:
 
         return False
 
-    leaves: List[Leaf] = []
+    new_line = line.clone()
     should_hug = False
     for idx, leaf in enumerate(line.leaves):
         new_leaf = leaf.clone()
@@ -139,18 +149,14 @@ def hug_power_op(line: Line, features: Collection[Feature]) -> Iterator[Line]:
         if should_hug:
             new_leaf.prefix = ""
 
-        leaves.append(new_leaf)
-
-    yield Line(
-        mode=line.mode,
-        depth=line.depth,
-        leaves=leaves,
-        comments=line.comments,
-        bracket_tracker=line.bracket_tracker,
-        inside_brackets=line.inside_brackets,
-        should_split_rhs=line.should_split_rhs,
-        magic_trailing_comma=line.magic_trailing_comma,
-    )
+        # We have to be careful to make a new line properly:
+        # - bracket related metadata must be maintained (handled by Line.append)
+        # - comments need to copied over, updating the leaf IDs they're attached to
+        new_line.append(new_leaf, preformatted=True)
+        for comment_leaf in line.comments_after(leaf):
+            new_line.append(comment_leaf, preformatted=True)
+
+    yield new_line
 
 
 class StringTransformer(ABC):
@@ -345,7 +351,7 @@ class StringMerger(StringTransformer, CustomSplitMapMixin):
 
     Requirements:
         (A) The line contains adjacent strings such that ALL of the validation checks
-        listed in StringMerger.__validate_msg(...)'s docstring pass.
+        listed in StringMerger._validate_msg(...)'s docstring pass.
             OR
         (B) The line contains a string which uses line continuation backslashes.
 
@@ -371,6 +377,8 @@ class StringMerger(StringTransformer, CustomSplitMapMixin):
                 and is_valid_index(i + 1)
                 and LL[i + 1].type == token.STRING
             ):
+                if is_part_of_annotation(leaf):
+                    return TErr("String is part of type annotation.")
                 return Ok(i)
 
             if leaf.type == token.STRING and "\\\n" in leaf.value:
@@ -448,7 +456,7 @@ class StringMerger(StringTransformer, CustomSplitMapMixin):
 
         Returns:
             Ok(new_line), if ALL of the validation checks found in
-            __validate_msg(...) pass.
+            _validate_msg(...) pass.
                 OR
             Err(CannotTransform), otherwise.
         """
@@ -547,6 +555,9 @@ class StringMerger(StringTransformer, CustomSplitMapMixin):
 
             next_str_idx += 1
 
+        # Take a note on the index of the non-STRING leaf.
+        non_string_idx = next_str_idx
+
         S_leaf = Leaf(token.STRING, S)
         if self.normalize_strings:
             S_leaf.value = normalize_string_quotes(S_leaf.value)
@@ -566,7 +577,18 @@ class StringMerger(StringTransformer, CustomSplitMapMixin):
         string_leaf = Leaf(token.STRING, S_leaf.value.replace(BREAK_MARK, ""))
 
         if atom_node is not None:
-            replace_child(atom_node, string_leaf)
+            # If not all children of the atom node are merged (this can happen
+            # when there is a standalone comment in the middle) ...
+            if non_string_idx - string_idx < len(atom_node.children):
+                # We need to replace the old STRING leaves with the new string leaf.
+                first_child_idx = LL[string_idx].remove()
+                for idx in range(string_idx + 1, non_string_idx):
+                    LL[idx].remove()
+                if first_child_idx is not None:
+                    atom_node.insert_child(first_child_idx, string_leaf)
+            else:
+                # Else replace the atom node with the new string leaf.
+                replace_child(atom_node, string_leaf)
 
         # Build the final line ('new_line') that this method will later return.
         new_line = line.clone()
@@ -588,7 +610,7 @@ class StringMerger(StringTransformer, CustomSplitMapMixin):
     def _validate_msg(line: Line, string_idx: int) -> TResult[None]:
         """Validate (M)erge (S)tring (G)roup
 
-        Transform-time string validation logic for __merge_string_group(...).
+        Transform-time string validation logic for _merge_string_group(...).
 
         Returns:
             * Ok(None), if ALL validation checks (listed below) pass.
@@ -602,6 +624,11 @@ class StringMerger(StringTransformer, CustomSplitMapMixin):
                 - The set of all string prefixes in the string group is of
                   length greater than one and is not equal to {"", "f"}.
                 - The string group consists of raw strings.
+                - The string group is stringified type annotations. We don't want to
+                  process stringified type annotations since pyright doesn't support
+                  them spanning multiple string values. (NOTE: mypy, pytype, pyre do
+                  support them, so we can change if pyright also gains support in the
+                  future. See https://github.com/microsoft/pyright/issues/4359.)
         """
         # We first check for "inner" stand-alone comments (i.e. stand-alone
         # comments that have a string leaf before them AND after them).
@@ -792,13 +819,7 @@ class StringParenStripper(StringTransformer):
 
         new_line = line.clone()
         new_line.comments = line.comments.copy()
-        try:
-            append_leaves(new_line, line, LL[: string_idx - 1])
-        except BracketMatchError:
-            # HACK: I believe there is currently a bug somewhere in
-            # right_hand_split() that is causing brackets to not be tracked
-            # properly by a shared BracketTracker.
-            append_leaves(new_line, line, LL[: string_idx - 1], preformatted=True)
+        append_leaves(new_line, line, LL[: string_idx - 1])
 
         string_leaf = Leaf(token.STRING, LL[string_idx].value)
         LL[string_idx - 1].remove()
@@ -1023,6 +1044,37 @@ class BaseStringSplitter(StringTransformer):
         max_string_length = self.line_length - offset
         return max_string_length
 
+    @staticmethod
+    def _prefer_paren_wrap_match(LL: List[Leaf]) -> Optional[int]:
+        """
+        Returns:
+            string_idx such that @LL[string_idx] is equal to our target (i.e.
+            matched) string, if this line matches the "prefer paren wrap" statement
+            requirements listed in the 'Requirements' section of the StringParenWrapper
+            class's docstring.
+                OR
+            None, otherwise.
+        """
+        # The line must start with a string.
+        if LL[0].type != token.STRING:
+            return None
+
+        # If the string is surrounded by commas (or is the first/last child)...
+        prev_sibling = LL[0].prev_sibling
+        next_sibling = LL[0].next_sibling
+        if not prev_sibling and not next_sibling and parent_type(LL[0]) == syms.atom:
+            # If it's an atom string, we need to check the parent atom's siblings.
+            parent = LL[0].parent
+            assert parent is not None  # For type checkers.
+            prev_sibling = parent.prev_sibling
+            next_sibling = parent.next_sibling
+        if (not prev_sibling or prev_sibling.type == token.COMMA) and (
+            not next_sibling or next_sibling.type == token.COMMA
+        ):
+            return 0
+
+        return None
+
 
 def iter_fexpr_spans(s: str) -> Iterator[Tuple[int, int]]:
     """
@@ -1118,6 +1170,9 @@ class StringSplitter(BaseStringSplitter, CustomSplitMapMixin):
     def do_splitter_match(self, line: Line) -> TMatchResult:
         LL = line.leaves
 
+        if self._prefer_paren_wrap_match(LL) is not None:
+            return TErr("Line needs to be wrapped in parens first.")
+
         is_valid_index = is_valid_index_factory(LL)
 
         idx = 0
@@ -1188,7 +1243,7 @@ class StringSplitter(BaseStringSplitter, CustomSplitMapMixin):
 
         string_op_leaves = self._get_string_operator_leaves(LL)
         string_op_leaves_length = (
-            sum([len(str(prefix_leaf)) for prefix_leaf in string_op_leaves]) + 1
+            sum(len(str(prefix_leaf)) for prefix_leaf in string_op_leaves) + 1
             if string_op_leaves
             else 0
         )
@@ -1305,9 +1360,14 @@ class StringSplitter(BaseStringSplitter, CustomSplitMapMixin):
             # prefix, and the current custom split did NOT originally use a
             # prefix...
             if (
-                next_value != self._normalize_f_string(next_value, prefix)
-                and use_custom_breakpoints
+                use_custom_breakpoints
                 and not csplit.has_prefix
+                and (
+                    # `next_value == prefix + QUOTE` happens when the custom
+                    # split is an empty string.
+                    next_value == prefix + QUOTE
+                    or next_value != self._normalize_f_string(next_value, prefix)
+                )
             ):
                 # Then `csplit.break_idx` will be off by one after removing
                 # the 'f' prefix.
@@ -1563,8 +1623,7 @@ class StringSplitter(BaseStringSplitter, CustomSplitMapMixin):
 
 class StringParenWrapper(BaseStringSplitter, CustomSplitMapMixin):
     """
-    StringTransformer that splits non-"atom" strings (i.e. strings that do not
-    exist on lines by themselves).
+    StringTransformer that wraps strings in parens and then splits at the LPAR.
 
     Requirements:
         All of the requirements listed in BaseStringSplitter's docstring in
@@ -1584,6 +1643,12 @@ class StringParenWrapper(BaseStringSplitter, CustomSplitMapMixin):
             OR
         * The line is a dictionary key assignment where some valid key is being
         assigned the value of some string.
+            OR
+        * The line is an lambda expression and the value is a string.
+            OR
+        * The line starts with an "atom" string that prefers to be wrapped in
+        parens. It's preferred to be wrapped when the string is surrounded by
+        commas (or is the first/last child).
 
     Transformations:
         The chosen string is wrapped in parentheses and then split at the LPAR.
@@ -1608,6 +1673,9 @@ class StringParenWrapper(BaseStringSplitter, CustomSplitMapMixin):
         changed such that it no longer needs to be given its own line,
         StringParenWrapper relies on StringParenStripper to clean up the
         parentheses it created.
+
+        For "atom" strings that prefers to be wrapped in parens, it requires
+        StringSplitter to hold the split until the string is wrapped in parens.
     """
 
     def do_splitter_match(self, line: Line) -> TMatchResult:
@@ -1623,7 +1691,8 @@ class StringParenWrapper(BaseStringSplitter, CustomSplitMapMixin):
             or self._else_match(LL)
             or self._assert_match(LL)
             or self._assign_match(LL)
-            or self._dict_match(LL)
+            or self._dict_or_lambda_match(LL)
+            or self._prefer_paren_wrap_match(LL)
         )
 
         if string_idx is not None:
@@ -1780,23 +1849,24 @@ class StringParenWrapper(BaseStringSplitter, CustomSplitMapMixin):
         return None
 
     @staticmethod
-    def _dict_match(LL: List[Leaf]) -> Optional[int]:
+    def _dict_or_lambda_match(LL: List[Leaf]) -> Optional[int]:
         """
         Returns:
             string_idx such that @LL[string_idx] is equal to our target (i.e.
             matched) string, if this line matches the dictionary key assignment
-            statement requirements listed in the 'Requirements' section of this
-            classes' docstring.
+            statement or lambda expression requirements listed in the
+            'Requirements' section of this classes' docstring.
                 OR
             None, otherwise.
         """
-        # If this line is apart of a dictionary key assignment...
-        if syms.dictsetmaker in [parent_type(LL[0]), parent_type(LL[0].parent)]:
+        # If this line is a part of a dictionary key assignment or lambda expression...
+        parent_types = [parent_type(LL[0]), parent_type(LL[0].parent)]
+        if syms.dictsetmaker in parent_types or syms.lambdef in parent_types:
             is_valid_index = is_valid_index_factory(LL)
 
             for i, leaf in enumerate(LL):
-                # We MUST find a colon...
-                if leaf.type == token.COLON:
+                # We MUST find a colon, it can either be dict's or lambda's colon...
+                if leaf.type == token.COLON and i < len(LL) - 1:
                     idx = i + 2 if is_empty_par(LL[i + 1]) else i + 1
 
                     # That colon MUST be followed by a string...
@@ -1890,6 +1960,25 @@ class StringParenWrapper(BaseStringSplitter, CustomSplitMapMixin):
                     f" (left_leaves={left_leaves}, right_leaves={right_leaves})"
                 )
                 old_rpar_leaf = right_leaves.pop()
+            elif right_leaves and right_leaves[-1].type == token.RPAR:
+                # Special case for lambda expressions as dict's value, e.g.:
+                #     my_dict = {
+                #        "key": lambda x: f"formatted: {x},
+                #     }
+                # After wrapping the dict's value with parentheses, the string is
+                # followed by a RPAR but its opening bracket is lambda's, not
+                # the string's:
+                #        "key": (lambda x: f"formatted: {x}),
+                opening_bracket = right_leaves[-1].opening_bracket
+                if opening_bracket is not None and opening_bracket in left_leaves:
+                    index = left_leaves.index(opening_bracket)
+                    if (
+                        index > 0
+                        and index < len(left_leaves) - 1
+                        and left_leaves[index - 1].type == token.COLON
+                        and left_leaves[index + 1].value == "lambda"
+                    ):
+                        right_leaves.pop()
 
             append_leaves(string_line, line, right_leaves)