]> 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:

Delay worker count determination
[etc/vim.git] / src / black / trans.py
index a4d1e6fbc7990f3813dfcc469a906af519252eb7..9e0284cefe3735b04b4b9ae472e4797d1ea4cb83 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 regex as re  # We need recursive patterns here (?R)
 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 Final
+    from typing_extensions import Final, Literal
 else:
-    from typing import Final
+    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,
+    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):
@@ -71,6 +81,84 @@ def TErr(err_msg: str) -> Err[CannotTransform]:
     return Err(cant_transform)
 
 
+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
+
+    new_line = line.clone()
+    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 = ""
+
+        # 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):
     """
     An implementation of the Transformer protocol that relies on its
@@ -283,7 +371,7 @@ class StringMerger(StringTransformer, CustomSplitMapMixin):
 
         is_valid_index = is_valid_index_factory(LL)
 
-        for (i, leaf) in enumerate(LL):
+        for i, leaf in enumerate(LL):
             if (
                 leaf.type == token.STRING
                 and is_valid_index(i + 1)
@@ -453,7 +541,7 @@ class StringMerger(StringTransformer, CustomSplitMapMixin):
             # 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)
 
@@ -465,6 +553,9 @@ 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)
@@ -484,11 +575,22 @@ 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()
-        for (i, leaf) in enumerate(LL):
+        for i, leaf in enumerate(LL):
             if i == string_idx:
                 new_line.append(string_leaf)
 
@@ -609,7 +711,7 @@ class StringParenStripper(StringTransformer):
 
         is_valid_index = is_valid_index_factory(LL)
 
-        for (idx, leaf) in enumerate(LL):
+        for idx, leaf in enumerate(LL):
             # Should be a string...
             if leaf.type != token.STRING:
                 continue
@@ -942,6 +1044,57 @@ class BaseStringSplitter(StringTransformer):
         return max_string_length
 
 
+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
@@ -981,17 +1134,6 @@ class StringSplitter(BaseStringSplitter, CustomSplitMapMixin):
     """
 
     MIN_SUBSTR_SIZE: Final = 6
-    # Matches an "f-expression" (e.g. {var}) that might be found in an f-string.
-    RE_FEXPR: Final = r"""
-    (?<!\{) (?:\{\{)* \{ (?!\{)
-        (?:
-            [^\{\}]
-            | \{\{
-            | \}\}
-            | (?R)
-        )+
-    \}
-    """
 
     def do_splitter_match(self, line: Line) -> TMatchResult:
         LL = line.leaves
@@ -1058,8 +1200,8 @@ class StringSplitter(BaseStringSplitter, CustomSplitMapMixin):
         # 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
@@ -1299,9 +1441,7 @@ class StringSplitter(BaseStringSplitter, CustomSplitMapMixin):
         """
         if "f" not in get_string_prefix(string).lower():
             return
-
-        for match in re.finditer(self.RE_FEXPR, string, re.VERBOSE):
-            yield match.span()
+        yield from iter_fexpr_spans(string)
 
     def _get_illegal_split_indices(self, string: str) -> Set[Index]:
         illegal_indices: Set[Index] = set()
@@ -1417,7 +1557,7 @@ class StringSplitter(BaseStringSplitter, CustomSplitMapMixin):
         """
         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) :]
@@ -1593,7 +1733,7 @@ class StringParenWrapper(BaseStringSplitter, CustomSplitMapMixin):
         if parent_type(LL[0]) == syms.assert_stmt and LL[0].value == "assert":
             is_valid_index = is_valid_index_factory(LL)
 
-            for (i, leaf) in enumerate(LL):
+            for i, leaf in enumerate(LL):
                 # We MUST find a comma...
                 if leaf.type == token.COMMA:
                     idx = i + 2 if is_empty_par(LL[i + 1]) else i + 1
@@ -1631,7 +1771,7 @@ class StringParenWrapper(BaseStringSplitter, CustomSplitMapMixin):
         ):
             is_valid_index = is_valid_index_factory(LL)
 
-            for (i, leaf) in enumerate(LL):
+            for i, leaf in enumerate(LL):
                 # We MUST find either an '=' or '+=' symbol...
                 if leaf.type in [token.EQUAL, token.PLUSEQUAL]:
                     idx = i + 2 if is_empty_par(LL[i + 1]) else i + 1
@@ -1674,7 +1814,7 @@ class StringParenWrapper(BaseStringSplitter, CustomSplitMapMixin):
         if syms.dictsetmaker in [parent_type(LL[0]), parent_type(LL[0].parent)]:
             is_valid_index = is_valid_index_factory(LL)
 
-            for (i, leaf) in enumerate(LL):
+            for i, leaf in enumerate(LL):
                 # We MUST find a colon...
                 if leaf.type == token.COLON:
                     idx = i + 2 if is_empty_par(LL[i + 1]) else i + 1