]> 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:

Use parentheses on method access on float and int literals (#2799)
[etc/vim.git] / src / black / trans.py
index d56205d4598bef11e27e4d0486ba1f1e2d95f629..74d052fe2dcbd0854d11bee16417777e28ed26eb 100644 (file)
@@ -4,10 +4,11 @@ String transformers that can split and merge strings.
 from abc import ABC, abstractmethod
 from collections import defaultdict
 from dataclasses import dataclass
 from abc import ABC, abstractmethod
 from collections import defaultdict
 from dataclasses import dataclass
-import regex as re
+import re
 from typing import (
     Any,
     Callable,
 from typing import (
     Any,
     Callable,
+    ClassVar,
     Collection,
     Dict,
     Iterable,
     Collection,
     Dict,
     Iterable,
@@ -15,10 +16,19 @@ from typing import (
     List,
     Optional,
     Sequence,
     List,
     Optional,
     Sequence,
+    Set,
     Tuple,
     TypeVar,
     Union,
 )
     Tuple,
     TypeVar,
     Union,
 )
+import sys
+
+if sys.version_info < (3, 8):
+    from typing_extensions import Literal, Final
+else:
+    from typing import Literal, Final
+
+from mypy_extensions import trait
 
 from black.rusty import Result, Ok, Err
 
 
 from black.rusty import Result, Ok, Err
 
@@ -61,7 +71,88 @@ def TErr(err_msg: str) -> Err[CannotTransform]:
     return Err(cant_transform)
 
 
     return Err(cant_transform)
 
 
-@dataclass  # type: ignore
+def hug_power_op(line: Line, features: Collection[Feature]) -> Iterator[Line]:
+    """A transformer which normalizes spacing around power operators."""
+
+    # Performance optimization to avoid unnecessary Leaf clones and other ops.
+    for leaf in line.leaves:
+        if leaf.type == token.DOUBLESTAR:
+            break
+    else:
+        raise CannotTransform("No doublestar token was found in the line.")
+
+    def is_simple_lookup(index: int, step: Literal[1, -1]) -> bool:
+        # Brackets and parentheses indicate calls, subscripts, etc. ...
+        # basically stuff that doesn't count as "simple". Only a NAME lookup
+        # or dotted lookup (eg. NAME.NAME) is OK.
+        if step == -1:
+            disallowed = {token.RPAR, token.RSQB}
+        else:
+            disallowed = {token.LPAR, token.LSQB}
+
+        while 0 <= index < len(line.leaves):
+            current = line.leaves[index]
+            if current.type in disallowed:
+                return False
+            if current.type not in {token.NAME, token.DOT} or current.value == "for":
+                # If the current token isn't disallowed, we'll assume this is simple as
+                # only the disallowed tokens are semantically attached to this lookup
+                # expression we're checking. Also, stop early if we hit the 'for' bit
+                # of a comprehension.
+                return True
+
+            index += step
+
+        return True
+
+    def is_simple_operand(index: int, kind: Literal["base", "exponent"]) -> bool:
+        # An operand is considered "simple" if's a NAME, a numeric CONSTANT, a simple
+        # lookup (see above), with or without a preceding unary operator.
+        start = line.leaves[index]
+        if start.type in {token.NAME, token.NUMBER}:
+            return is_simple_lookup(index, step=(1 if kind == "exponent" else -1))
+
+        if start.type in {token.PLUS, token.MINUS, token.TILDE}:
+            if line.leaves[index + 1].type in {token.NAME, token.NUMBER}:
+                # step is always one as bases with a preceding unary op will be checked
+                # for simplicity starting from the next token (so it'll hit the check
+                # above).
+                return is_simple_lookup(index + 1, step=1)
+
+        return False
+
+    leaves: List[Leaf] = []
+    should_hug = False
+    for idx, leaf in enumerate(line.leaves):
+        new_leaf = leaf.clone()
+        if should_hug:
+            new_leaf.prefix = ""
+            should_hug = False
+
+        should_hug = (
+            (0 < idx < len(line.leaves) - 1)
+            and leaf.type == token.DOUBLESTAR
+            and is_simple_operand(idx - 1, kind="base")
+            and line.leaves[idx - 1].value != "lambda"
+            and is_simple_operand(idx + 1, kind="exponent")
+        )
+        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,
+    )
+
+
 class StringTransformer(ABC):
     """
     An implementation of the Transformer protocol that relies on its
 class StringTransformer(ABC):
     """
     An implementation of the Transformer protocol that relies on its
@@ -89,9 +180,13 @@ class StringTransformer(ABC):
         as much as possible.
     """
 
         as much as possible.
     """
 
-    line_length: int
-    normalize_strings: bool
-    __name__ = "StringTransformer"
+    __name__: Final = "StringTransformer"
+
+    # Ideally this would be a dataclass, but unfortunately mypyc breaks when used with
+    # `abc.ABC`.
+    def __init__(self, line_length: int, normalize_strings: bool) -> None:
+        self.line_length = line_length
+        self.normalize_strings = normalize_strings
 
     @abstractmethod
     def do_match(self, line: Line) -> TMatchResult:
 
     @abstractmethod
     def do_match(self, line: Line) -> TMatchResult:
@@ -183,6 +278,7 @@ class CustomSplit:
     break_idx: int
 
 
     break_idx: int
 
 
+@trait
 class CustomSplitMapMixin:
     """
     This mixin class is used to map merged strings to a sequence of
 class CustomSplitMapMixin:
     """
     This mixin class is used to map merged strings to a sequence of
@@ -190,8 +286,10 @@ class CustomSplitMapMixin:
     the resultant substrings go over the configured max line length.
     """
 
     the resultant substrings go over the configured max line length.
     """
 
-    _Key = Tuple[StringID, str]
-    _CUSTOM_SPLIT_MAP: Dict[_Key, Tuple[CustomSplit, ...]] = defaultdict(tuple)
+    _Key: ClassVar = Tuple[StringID, str]
+    _CUSTOM_SPLIT_MAP: ClassVar[Dict[_Key, Tuple[CustomSplit, ...]]] = defaultdict(
+        tuple
+    )
 
     @staticmethod
     def _get_key(string: str) -> "CustomSplitMapMixin._Key":
 
     @staticmethod
     def _get_key(string: str) -> "CustomSplitMapMixin._Key":
@@ -242,7 +340,7 @@ class CustomSplitMapMixin:
         return key in self._CUSTOM_SPLIT_MAP
 
 
         return key in self._CUSTOM_SPLIT_MAP
 
 
-class StringMerger(CustomSplitMapMixin, StringTransformer):
+class StringMerger(StringTransformer, CustomSplitMapMixin):
     """StringTransformer that merges strings together.
 
     Requirements:
     """StringTransformer that merges strings together.
 
     Requirements:
@@ -411,7 +509,7 @@ class StringMerger(CustomSplitMapMixin, StringTransformer):
             and is_valid_index(next_str_idx)
             and LL[next_str_idx].type == token.STRING
         ):
             and is_valid_index(next_str_idx)
             and LL[next_str_idx].type == token.STRING
         ):
-            prefix = get_string_prefix(LL[next_str_idx].value)
+            prefix = get_string_prefix(LL[next_str_idx].value).lower()
             next_str_idx += 1
 
         # The next loop merges the string group. The final string will be
             next_str_idx += 1
 
         # The next loop merges the string group. The final string will be
@@ -431,13 +529,13 @@ class StringMerger(CustomSplitMapMixin, StringTransformer):
             num_of_strings += 1
 
             SS = LL[next_str_idx].value
             num_of_strings += 1
 
             SS = LL[next_str_idx].value
-            next_prefix = get_string_prefix(SS)
+            next_prefix = get_string_prefix(SS).lower()
 
             # If this is an f-string group but this substring is not prefixed
             # with 'f'...
             if "f" in prefix and "f" not in next_prefix:
                 # Then we must escape any braces contained in this substring.
 
             # If this is an f-string group but this substring is not prefixed
             # with 'f'...
             if "f" in prefix and "f" not in next_prefix:
                 # Then we must escape any braces contained in this substring.
-                SS = re.subf(r"(\{|\})", "{1}{1}", SS)
+                SS = re.sub(r"(\{|\})", r"\1\1", SS)
 
             NSS = make_naked(SS, next_prefix)
 
 
             NSS = make_naked(SS, next_prefix)
 
@@ -541,7 +639,7 @@ class StringMerger(CustomSplitMapMixin, StringTransformer):
                 return TErr("StringMerger does NOT merge multiline strings.")
 
             num_of_strings += 1
                 return TErr("StringMerger does NOT merge multiline strings.")
 
             num_of_strings += 1
-            prefix = get_string_prefix(leaf.value)
+            prefix = get_string_prefix(leaf.value).lower()
             if "r" in prefix:
                 return TErr("StringMerger does NOT merge raw strings.")
 
             if "r" in prefix:
                 return TErr("StringMerger does NOT merge raw strings.")
 
@@ -738,7 +836,7 @@ class BaseStringSplitter(StringTransformer):
         * The target string is not a multiline (i.e. triple-quote) string.
     """
 
         * The target string is not a multiline (i.e. triple-quote) string.
     """
 
-    STRING_OPERATORS = [
+    STRING_OPERATORS: Final = [
         token.EQEQUAL,
         token.GREATER,
         token.GREATEREQUAL,
         token.EQEQUAL,
         token.GREATER,
         token.GREATEREQUAL,
@@ -926,7 +1024,58 @@ class BaseStringSplitter(StringTransformer):
         return max_string_length
 
 
         return max_string_length
 
 
-class StringSplitter(CustomSplitMapMixin, BaseStringSplitter):
+def iter_fexpr_spans(s: str) -> Iterator[Tuple[int, int]]:
+    """
+    Yields spans corresponding to expressions in a given f-string.
+    Spans are half-open ranges (left inclusive, right exclusive).
+    Assumes the input string is a valid f-string, but will not crash if the input
+    string is invalid.
+    """
+    stack: List[int] = []  # our curly paren stack
+    i = 0
+    while i < len(s):
+        if s[i] == "{":
+            # if we're in a string part of the f-string, ignore escaped curly braces
+            if not stack and i + 1 < len(s) and s[i + 1] == "{":
+                i += 2
+                continue
+            stack.append(i)
+            i += 1
+            continue
+
+        if s[i] == "}":
+            if not stack:
+                i += 1
+                continue
+            j = stack.pop()
+            # we've made it back out of the expression! yield the span
+            if not stack:
+                yield (j, i + 1)
+            i += 1
+            continue
+
+        # if we're in an expression part of the f-string, fast forward through strings
+        # note that backslashes are not legal in the expression portion of f-strings
+        if stack:
+            delim = None
+            if s[i : i + 3] in ("'''", '"""'):
+                delim = s[i : i + 3]
+            elif s[i] in ("'", '"'):
+                delim = s[i]
+            if delim:
+                i += len(delim)
+                while i < len(s) and s[i : i + len(delim)] != delim:
+                    i += 1
+                i += len(delim)
+                continue
+        i += 1
+
+
+def fstring_contains_expr(s: str) -> bool:
+    return any(iter_fexpr_spans(s))
+
+
+class StringSplitter(BaseStringSplitter, CustomSplitMapMixin):
     """
     StringTransformer that splits "atom" strings (i.e. strings which exist on
     lines by themselves).
     """
     StringTransformer that splits "atom" strings (i.e. strings which exist on
     lines by themselves).
@@ -964,18 +1113,7 @@ class StringSplitter(CustomSplitMapMixin, BaseStringSplitter):
         CustomSplit objects and add them to the custom split map.
     """
 
         CustomSplit objects and add them to the custom split map.
     """
 
-    MIN_SUBSTR_SIZE = 6
-    # Matches an "f-expression" (e.g. {var}) that might be found in an f-string.
-    RE_FEXPR = r"""
-    (?<!\{) (?:\{\{)* \{ (?!\{)
-        (?:
-            [^\{\}]
-            | \{\{
-            | \}\}
-            | (?R)
-        )+
-    \}
-    """
+    MIN_SUBSTR_SIZE: Final = 6
 
     def do_splitter_match(self, line: Line) -> TMatchResult:
         LL = line.leaves
 
     def do_splitter_match(self, line: Line) -> TMatchResult:
         LL = line.leaves
@@ -1036,14 +1174,14 @@ class StringSplitter(CustomSplitMapMixin, BaseStringSplitter):
         is_valid_index = is_valid_index_factory(LL)
         insert_str_child = insert_str_child_factory(LL[string_idx])
 
         is_valid_index = is_valid_index_factory(LL)
         insert_str_child = insert_str_child_factory(LL[string_idx])
 
-        prefix = get_string_prefix(LL[string_idx].value)
+        prefix = get_string_prefix(LL[string_idx].value).lower()
 
         # We MAY choose to drop the 'f' prefix from substrings that don't
         # contain any f-expressions, but ONLY if the original f-string
         # contains at least one f-expression. Otherwise, we will alter the AST
         # of the program.
 
         # We MAY choose to drop the 'f' prefix from substrings that don't
         # contain any f-expressions, but ONLY if the original f-string
         # contains at least one f-expression. Otherwise, we will alter the AST
         # of the program.
-        drop_pointless_f_prefix = ("f" in prefix) and re.search(
-            self.RE_FEXPR, LL[string_idx].value, re.VERBOSE
+        drop_pointless_f_prefix = ("f" in prefix) and fstring_contains_expr(
+            LL[string_idx].value
         )
 
         first_string_line = True
         )
 
         first_string_line = True
@@ -1151,21 +1289,32 @@ class StringSplitter(CustomSplitMapMixin, BaseStringSplitter):
 
             # --- Construct `next_value`
             next_value = rest_value[:break_idx] + QUOTE
 
             # --- Construct `next_value`
             next_value = rest_value[:break_idx] + QUOTE
+
+            # HACK: The following 'if' statement is a hack to fix the custom
+            # breakpoint index in the case of either: (a) substrings that were
+            # f-strings but will have the 'f' prefix removed OR (b) substrings
+            # that were not f-strings but will now become f-strings because of
+            # redundant use of the 'f' prefix (i.e. none of the substrings
+            # contain f-expressions but one or more of them had the 'f' prefix
+            # anyway; in which case, we will prepend 'f' to _all_ substrings).
+            #
+            # There is probably a better way to accomplish what is being done
+            # here...
+            #
+            # If this substring is an f-string, we _could_ remove the 'f'
+            # prefix, and the current custom split did NOT originally use a
+            # prefix...
             if (
             if (
-                # Are we allowed to try to drop a pointless 'f' prefix?
-                drop_pointless_f_prefix
-                # If we are, will we be successful?
-                and next_value != self._normalize_f_string(next_value, prefix)
+                next_value != self._normalize_f_string(next_value, prefix)
+                and use_custom_breakpoints
+                and not csplit.has_prefix
             ):
             ):
-                # If the current custom split did NOT originally use a prefix,
-                # then `csplit.break_idx` will be off by one after removing
+                # Then `csplit.break_idx` will be off by one after removing
                 # the 'f' prefix.
                 # the 'f' prefix.
-                break_idx = (
-                    break_idx + 1
-                    if use_custom_breakpoints and not csplit.has_prefix
-                    else break_idx
-                )
+                break_idx += 1
                 next_value = rest_value[:break_idx] + QUOTE
                 next_value = rest_value[:break_idx] + QUOTE
+
+            if drop_pointless_f_prefix:
                 next_value = self._normalize_f_string(next_value, prefix)
 
             # --- Construct `next_leaf`
                 next_value = self._normalize_f_string(next_value, prefix)
 
             # --- Construct `next_leaf`
@@ -1232,6 +1381,59 @@ class StringSplitter(CustomSplitMapMixin, BaseStringSplitter):
             last_line.comments = line.comments.copy()
             yield Ok(last_line)
 
             last_line.comments = line.comments.copy()
             yield Ok(last_line)
 
+    def _iter_nameescape_slices(self, string: str) -> Iterator[Tuple[Index, Index]]:
+        """
+        Yields:
+            All ranges of @string which, if @string were to be split there,
+            would result in the splitting of an \\N{...} expression (which is NOT
+            allowed).
+        """
+        # True - the previous backslash was unescaped
+        # False - the previous backslash was escaped *or* there was no backslash
+        previous_was_unescaped_backslash = False
+        it = iter(enumerate(string))
+        for idx, c in it:
+            if c == "\\":
+                previous_was_unescaped_backslash = not previous_was_unescaped_backslash
+                continue
+            if not previous_was_unescaped_backslash or c != "N":
+                previous_was_unescaped_backslash = False
+                continue
+            previous_was_unescaped_backslash = False
+
+            begin = idx - 1  # the position of backslash before \N{...}
+            for idx, c in it:
+                if c == "}":
+                    end = idx
+                    break
+            else:
+                # malformed nameescape expression?
+                # should have been detected by AST parsing earlier...
+                raise RuntimeError(f"{self.__class__.__name__} LOGIC ERROR!")
+            yield begin, end
+
+    def _iter_fexpr_slices(self, string: str) -> Iterator[Tuple[Index, Index]]:
+        """
+        Yields:
+            All ranges of @string which, if @string were to be split there,
+            would result in the splitting of an f-expression (which is NOT
+            allowed).
+        """
+        if "f" not in get_string_prefix(string).lower():
+            return
+        yield from iter_fexpr_spans(string)
+
+    def _get_illegal_split_indices(self, string: str) -> Set[Index]:
+        illegal_indices: Set[Index] = set()
+        iterators = [
+            self._iter_fexpr_slices(string),
+            self._iter_nameescape_slices(string),
+        ]
+        for it in iterators:
+            for begin, end in it:
+                illegal_indices.update(range(begin, end + 1))
+        return illegal_indices
+
     def _get_break_idx(self, string: str, max_break_idx: int) -> Optional[int]:
         """
         This method contains the algorithm that StringSplitter uses to
     def _get_break_idx(self, string: str, max_break_idx: int) -> Optional[int]:
         """
         This method contains the algorithm that StringSplitter uses to
@@ -1261,40 +1463,15 @@ class StringSplitter(CustomSplitMapMixin, BaseStringSplitter):
         assert is_valid_index(max_break_idx)
         assert_is_leaf_string(string)
 
         assert is_valid_index(max_break_idx)
         assert_is_leaf_string(string)
 
-        _fexpr_slices: Optional[List[Tuple[Index, Index]]] = None
-
-        def fexpr_slices() -> Iterator[Tuple[Index, Index]]:
-            """
-            Yields:
-                All ranges of @string which, if @string were to be split there,
-                would result in the splitting of an f-expression (which is NOT
-                allowed).
-            """
-            nonlocal _fexpr_slices
-
-            if _fexpr_slices is None:
-                _fexpr_slices = []
-                for match in re.finditer(self.RE_FEXPR, string, re.VERBOSE):
-                    _fexpr_slices.append(match.span())
-
-            yield from _fexpr_slices
-
-        is_fstring = "f" in get_string_prefix(string)
+        _illegal_split_indices = self._get_illegal_split_indices(string)
 
 
-        def breaks_fstring_expression(i: Index) -> bool:
+        def breaks_unsplittable_expression(i: Index) -> bool:
             """
             Returns:
                 True iff returning @i would result in the splitting of an
             """
             Returns:
                 True iff returning @i would result in the splitting of an
-                f-expression (which is NOT allowed).
+                unsplittable expression (which is NOT allowed).
             """
             """
-            if not is_fstring:
-                return False
-
-            for (start, end) in fexpr_slices():
-                if start <= i < end:
-                    return True
-
-            return False
+            return i in _illegal_split_indices
 
         def passes_all_checks(i: Index) -> bool:
             """
 
         def passes_all_checks(i: Index) -> bool:
             """
@@ -1318,7 +1495,7 @@ class StringSplitter(CustomSplitMapMixin, BaseStringSplitter):
                 is_space
                 and is_not_escaped
                 and is_big_enough
                 is_space
                 and is_not_escaped
                 and is_big_enough
-                and not breaks_fstring_expression(i)
+                and not breaks_unsplittable_expression(i)
             )
 
         # First, we check all indices BELOW @max_break_idx.
             )
 
         # First, we check all indices BELOW @max_break_idx.
@@ -1360,7 +1537,7 @@ class StringSplitter(CustomSplitMapMixin, BaseStringSplitter):
         """
         assert_is_leaf_string(string)
 
         """
         assert_is_leaf_string(string)
 
-        if "f" in prefix and not re.search(self.RE_FEXPR, string, re.VERBOSE):
+        if "f" in prefix and not fstring_contains_expr(string):
             new_prefix = prefix.replace("f", "")
 
             temp = string[len(prefix) :]
             new_prefix = prefix.replace("f", "")
 
             temp = string[len(prefix) :]
@@ -1384,7 +1561,7 @@ class StringSplitter(CustomSplitMapMixin, BaseStringSplitter):
         return string_op_leaves
 
 
         return string_op_leaves
 
 
-class StringParenWrapper(CustomSplitMapMixin, BaseStringSplitter):
+class StringParenWrapper(BaseStringSplitter, CustomSplitMapMixin):
     """
     StringTransformer that splits non-"atom" strings (i.e. strings that do not
     exist on lines by themselves).
     """
     StringTransformer that splits non-"atom" strings (i.e. strings that do not
     exist on lines by themselves).
@@ -1769,20 +1946,20 @@ class StringParser:
         ```
     """
 
         ```
     """
 
-    DEFAULT_TOKEN = -1
+    DEFAULT_TOKEN: Final = 20210605
 
     # String Parser States
 
     # String Parser States
-    START = 1
-    DOT = 2
-    NAME = 3
-    PERCENT = 4
-    SINGLE_FMT_ARG = 5
-    LPAR = 6
-    RPAR = 7
-    DONE = 8
+    START: Final = 1
+    DOT: Final = 2
+    NAME: Final = 3
+    PERCENT: Final = 4
+    SINGLE_FMT_ARG: Final = 5
+    LPAR: Final = 6
+    RPAR: Final = 7
+    DONE: Final = 8
 
     # Lookup Table for Next State
 
     # Lookup Table for Next State
-    _goto: Dict[Tuple[ParserState, NodeType], ParserState] = {
+    _goto: Final[Dict[Tuple[ParserState, NodeType], ParserState]] = {
         # A string trailer may start with '.' OR '%'.
         (START, token.DOT): DOT,
         (START, token.PERCENT): PERCENT,
         # A string trailer may start with '.' OR '%'.
         (START, token.DOT): DOT,
         (START, token.PERCENT): PERCENT,