Update CHANGES.md for 21.6b0 release (#2325)

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

diff --git a/src/black/trans.py b/src/black/trans.py
index bc6e93b01b40706c4f19d8ee053664c105812712..023dcd3618a5c0edb6b6f5e7e91e09bb6d3add30 100644 (file)
--- a/src/black/trans.py
+++ b/src/black/trans.py
@@ -15,6 +15,7 @@ from typing import (
     List,
     Optional,
     Sequence,
     List,
     Optional,
     Sequence,

+    Set,

     Tuple,
     TypeVar,
     Union,
     Tuple,
     TypeVar,
     Union,


@@ -411,7 +412,7 @@ 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


@@ -431,7 +432,7 @@ 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 this is an f-string group but this substring is not prefixed
             # with 'f'...


@@ -541,7 +542,7 @@ 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.")
 


@@ -867,7 +868,7 @@ class BaseStringSplitter(StringTransformer):
                 # WMA4 a space, a comma, and a closing bracket [e.g. `), STRING`].
                 offset += 3
 
                 # WMA4 a space, a comma, and a closing bracket [e.g. `), STRING`].
                 offset += 3
 

-            if P.type in [token.COLON, token.EQUAL, token.NAME]:
+            if P.type in [token.COLON, token.EQUAL, token.PLUSEQUAL, token.NAME]:

                 # This conditional branch is meant to handle dictionary keys,
                 # variable assignments, 'return STRING' statement lines, and
                 # 'else STRING' ternary expression lines.
                 # This conditional branch is meant to handle dictionary keys,
                 # variable assignments, 'return STRING' statement lines, and
                 # 'else STRING' ternary expression lines.


@@ -1036,7 +1037,7 @@ 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
 
         # We MAY choose to drop the 'f' prefix from substrings that don't
         # contain any f-expressions, but ONLY if the original f-string


@@ -1151,21 +1152,32 @@ 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 +1244,61 @@ 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
+
+        for match in re.finditer(self.RE_FEXPR, string, re.VERBOSE):
+            yield match.span()
+
+    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 +1328,15 @@ 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
+        _illegal_split_indices = self._get_illegal_split_indices(string)

 
 

-        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)
-
-        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 +1360,7 @@ 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.