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

Always explode data structure literals
[etc/vim.git] / black.py
index e7a7aa8585836c72a5d24dfa35fda0089462578a..c48b8d102f58cf2dfd1c9f9db97cf75f6f405af6 100644 (file)
--- a/black.py
+++ b/black.py
@@ -282,32 +282,29 @@ async def schedule_formatting(
             manager = Manager()
             lock = manager.Lock()
         tasks = {
-            src: loop.run_in_executor(
+            loop.run_in_executor(
                 executor, format_file_in_place, src, line_length, fast, write_back, lock
-            )
-            for src in sources
+            ): src
+            for src in sorted(sources)
         }
-        _task_values = list(tasks.values())
+        pending: Iterable[asyncio.Task] = tasks.keys()
         try:
-            loop.add_signal_handler(signal.SIGINT, cancel, _task_values)
-            loop.add_signal_handler(signal.SIGTERM, cancel, _task_values)
+            loop.add_signal_handler(signal.SIGINT, cancel, pending)
+            loop.add_signal_handler(signal.SIGTERM, cancel, pending)
         except NotImplementedError:
             # There are no good alternatives for these on Windows
             pass
-        await asyncio.wait(_task_values)
-        for src, task in tasks.items():
-            if not task.done():
-                report.failed(src, "timed out, cancelling")
-                task.cancel()
-                cancelled.append(task)
-            elif task.cancelled():
-                cancelled.append(task)
-            elif task.exception():
-                report.failed(src, str(task.exception()))
-            else:
-                formatted.append(src)
-                report.done(src, Changed.YES if task.result() else Changed.NO)
-
+        while pending:
+            done, _ = await asyncio.wait(pending, return_when=asyncio.FIRST_COMPLETED)
+            for task in done:
+                src = tasks.pop(task)
+                if task.cancelled():
+                    cancelled.append(task)
+                elif task.exception():
+                    report.failed(src, str(task.exception()))
+                else:
+                    formatted.append(src)
+                    report.done(src, Changed.YES if task.result() else Changed.NO)
     if cancelled:
         await asyncio.gather(*cancelled, loop=loop, return_exceptions=True)
     if write_back == WriteBack.YES and formatted:
@@ -629,21 +626,22 @@ LOGIC_PRIORITY = 14
 STRING_PRIORITY = 12
 COMPARATOR_PRIORITY = 10
 MATH_PRIORITIES = {
-    token.VBAR: 8,
-    token.CIRCUMFLEX: 7,
-    token.AMPER: 6,
-    token.LEFTSHIFT: 5,
-    token.RIGHTSHIFT: 5,
-    token.PLUS: 4,
-    token.MINUS: 4,
-    token.STAR: 3,
-    token.SLASH: 3,
-    token.DOUBLESLASH: 3,
-    token.PERCENT: 3,
-    token.AT: 3,
-    token.TILDE: 2,
-    token.DOUBLESTAR: 1,
+    token.VBAR: 9,
+    token.CIRCUMFLEX: 8,
+    token.AMPER: 7,
+    token.LEFTSHIFT: 6,
+    token.RIGHTSHIFT: 6,
+    token.PLUS: 5,
+    token.MINUS: 5,
+    token.STAR: 4,
+    token.SLASH: 4,
+    token.DOUBLESLASH: 4,
+    token.PERCENT: 4,
+    token.AT: 4,
+    token.TILDE: 3,
+    token.DOUBLESTAR: 2,
 }
+DOT_PRIORITY = 1
 
 
 @dataclass
@@ -709,6 +707,17 @@ class BracketTracker:
         """
         return max(v for k, v in self.delimiters.items() if k not in exclude)
 
+    def delimiter_count_with_priority(self, priority: int = 0) -> int:
+        """Return the number of delimiters with the given `priority`.
+
+        If no `priority` is passed, defaults to max priority on the line.
+        """
+        if not self.delimiters:
+            return 0
+
+        priority = priority or self.max_delimiter_priority()
+        return sum(1 for p in self.delimiters.values() if p == priority)
+
     def maybe_increment_for_loop_variable(self, leaf: Leaf) -> bool:
         """In a for loop, or comprehension, the variables are often unpacks.
 
@@ -767,6 +776,7 @@ class Line:
     comments: List[Tuple[Index, Leaf]] = Factory(list)
     bracket_tracker: BracketTracker = Factory(BracketTracker)
     inside_brackets: bool = False
+    should_explode: bool = False
 
     def append(self, leaf: Leaf, preformatted: bool = False) -> None:
         """Add a new `leaf` to the end of the line.
@@ -840,10 +850,9 @@ class Line:
     @property
     def is_stub_class(self) -> bool:
         """Is this line a class definition with a body consisting only of "..."?"""
-        return (
-            self.is_class
-            and self.leaves[-3:] == [Leaf(token.DOT, ".") for _ in range(3)]
-        )
+        return self.is_class and self.leaves[-3:] == [
+            Leaf(token.DOT, ".") for _ in range(3)
+        ]
 
     @property
     def is_def(self) -> bool:
@@ -1307,7 +1316,7 @@ class LineGenerator(Visitor[Line]):
         The relevant Python language `keywords` for a given statement will be
         NAME leaves within it. This methods puts those on a separate line.
 
-        `parens` holds a set of string leaf values immeditely after which
+        `parens` holds a set of string leaf values immediately after which
         invisible parens should be put.
         """
         normalize_invisible_parens(node, parens_after=parens)
@@ -1464,10 +1473,11 @@ def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str:  # noqa C901
         return DOUBLESPACE
 
     assert p is not None, f"INTERNAL ERROR: hand-made leaf without parent: {leaf!r}"
-    if (
-        t == token.COLON
-        and p.type not in {syms.subscript, syms.subscriptlist, syms.sliceop}
-    ):
+    if t == token.COLON and p.type not in {
+        syms.subscript,
+        syms.subscriptlist,
+        syms.sliceop,
+    }:
         return NO
 
     prev = leaf.prev_sibling
@@ -1488,7 +1498,10 @@ def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str:  # noqa C901
         if prevp.type == token.EQUAL:
             if prevp.parent:
                 if prevp.parent.type in {
-                    syms.arglist, syms.argument, syms.parameters, syms.varargslist
+                    syms.arglist,
+                    syms.argument,
+                    syms.parameters,
+                    syms.varargslist,
                 }:
                     return NO
 
@@ -1641,10 +1654,10 @@ def whitespace(leaf: Leaf, *, complex_subscript: bool) -> str:  # noqa C901
 
             prevp_parent = prevp.parent
             assert prevp_parent is not None
-            if (
-                prevp.type == token.COLON
-                and prevp_parent.type in {syms.subscript, syms.sliceop}
-            ):
+            if prevp.type == token.COLON and prevp_parent.type in {
+                syms.subscript,
+                syms.sliceop,
+            }:
                 return NO
 
             elif prevp.type == token.EQUAL and prevp_parent.type == syms.argument:
@@ -1724,6 +1737,14 @@ def is_split_before_delimiter(leaf: Leaf, previous: Leaf = None) -> int:
         # Don't treat them as a delimiter.
         return 0
 
+    if (
+        leaf.type == token.DOT
+        and leaf.parent
+        and leaf.parent.type not in {syms.import_from, syms.dotted_name}
+        and (previous is None or previous.type != token.NAME)
+    ):
+        return DOT_PRIORITY
+
     if (
         leaf.type in MATH_OPERATORS
         and leaf.parent
@@ -1741,31 +1762,54 @@ def is_split_before_delimiter(leaf: Leaf, previous: Leaf = None) -> int:
     ):
         return STRING_PRIORITY
 
+    if leaf.type != token.NAME:
+        return 0
+
     if (
-        leaf.type == token.NAME
-        and leaf.value == "for"
+        leaf.value == "for"
         and leaf.parent
         and leaf.parent.type in {syms.comp_for, syms.old_comp_for}
     ):
         return COMPREHENSION_PRIORITY
 
     if (
-        leaf.type == token.NAME
-        and leaf.value == "if"
+        leaf.value == "if"
         and leaf.parent
         and leaf.parent.type in {syms.comp_if, syms.old_comp_if}
     ):
         return COMPREHENSION_PRIORITY
 
+    if leaf.value in {"if", "else"} and leaf.parent and leaf.parent.type == syms.test:
+        return TERNARY_PRIORITY
+
+    if leaf.value == "is":
+        return COMPARATOR_PRIORITY
+
     if (
-        leaf.type == token.NAME
-        and leaf.value in {"if", "else"}
+        leaf.value == "in"
         and leaf.parent
-        and leaf.parent.type == syms.test
+        and leaf.parent.type in {syms.comp_op, syms.comparison}
+        and not (
+            previous is not None
+            and previous.type == token.NAME
+            and previous.value == "not"
+        )
     ):
-        return TERNARY_PRIORITY
+        return COMPARATOR_PRIORITY
 
-    if leaf.type == token.NAME and leaf.value in LOGIC_OPERATORS and leaf.parent:
+    if (
+        leaf.value == "not"
+        and leaf.parent
+        and leaf.parent.type == syms.comp_op
+        and not (
+            previous is not None
+            and previous.type == token.NAME
+            and previous.value == "is"
+        )
+    ):
+        return COMPARATOR_PRIORITY
+
+    if leaf.value in LOGIC_OPERATORS and leaf.parent:
         return LOGIC_PRIORITY
 
     return 0
@@ -1865,15 +1909,15 @@ def split_line(
         return
 
     line_str = str(line).strip("\n")
-    if is_line_short_enough(line, line_length=line_length, line_str=line_str):
+    if not line.should_explode and is_line_short_enough(
+        line, line_length=line_length, line_str=line_str
+    ):
         yield line
         return
 
     split_funcs: List[SplitFunc]
     if line.is_def:
         split_funcs = [left_hand_split]
-    elif line.is_import:
-        split_funcs = [explode_split]
     else:
 
         def rhs(line: Line, py36: bool = False) -> Iterator[Line]:
@@ -2019,7 +2063,7 @@ def right_hand_split(
         and not line.is_import
     ):
         omit = {id(closing_bracket), *omit}
-        delimiter_count = len(body.bracket_tracker.delimiters)
+        delimiter_count = body.bracket_tracker.delimiter_count_with_priority()
         if (
             delimiter_count == 0
             or delimiter_count == 1
@@ -2036,6 +2080,7 @@ def right_hand_split(
 
     ensure_visible(opening_bracket)
     ensure_visible(closing_bracket)
+    body.should_explode = should_explode(body, opening_bracket)
     for result in (head, body, tail):
         if result:
             yield result
@@ -2094,14 +2139,16 @@ def delimiter_split(line: Line, py36: bool = False) -> Iterator[Line]:
     except IndexError:
         raise CannotSplit("Line empty")
 
-    delimiters = line.bracket_tracker.delimiters
+    bt = line.bracket_tracker
     try:
-        delimiter_priority = line.bracket_tracker.max_delimiter_priority(
-            exclude={id(last_leaf)}
-        )
+        delimiter_priority = bt.max_delimiter_priority(exclude={id(last_leaf)})
     except ValueError:
         raise CannotSplit("No delimiters found")
 
+    if delimiter_priority == DOT_PRIORITY:
+        if bt.delimiter_count_with_priority(delimiter_priority) == 1:
+            raise CannotSplit("Splitting a single attribute from its owner looks wrong")
+
     current_line = Line(depth=line.depth, inside_brackets=line.inside_brackets)
     lowest_depth = sys.maxsize
     trailing_comma_safe = True
@@ -2124,12 +2171,11 @@ def delimiter_split(line: Line, py36: bool = False) -> Iterator[Line]:
             yield from append_to_line(comment_after)
 
         lowest_depth = min(lowest_depth, leaf.bracket_depth)
-        if (
-            leaf.bracket_depth == lowest_depth
-            and is_vararg(leaf, within=VARARGS_PARENTS)
+        if leaf.bracket_depth == lowest_depth and is_vararg(
+            leaf, within=VARARGS_PARENTS
         ):
             trailing_comma_safe = trailing_comma_safe and py36
-        leaf_priority = delimiters.get(id(leaf))
+        leaf_priority = bt.delimiters.get(id(leaf))
         if leaf_priority == delimiter_priority:
             yield current_line
 
@@ -2174,26 +2220,6 @@ def standalone_comment_split(line: Line, py36: bool = False) -> Iterator[Line]:
         yield current_line
 
 
-def explode_split(
-    line: Line, py36: bool = False, omit: Collection[LeafID] = ()
-) -> Iterator[Line]:
-    """Split by rightmost bracket and immediately split contents by a delimiter."""
-    new_lines = list(right_hand_split(line, py36, omit))
-    if len(new_lines) != 3:
-        yield from new_lines
-        return
-
-    yield new_lines[0]
-
-    try:
-        yield from delimiter_split(new_lines[1], py36)
-
-    except CannotSplit:
-        yield new_lines[1]
-
-    yield new_lines[2]
-
-
 def is_import(leaf: Leaf) -> bool:
     """Return True if the given leaf starts an import statement."""
     p = leaf.parent
@@ -2323,7 +2349,7 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
                 rpar = Leaf(token.RPAR, ")")
                 index = child.remove() or 0
                 node.insert_child(index, Node(syms.atom, [lpar, child, rpar]))
-            else:
+            elif not (isinstance(child, Leaf) and is_multiline_string(child)):
                 # wrap child in invisible parentheses
                 lpar = Leaf(token.LPAR, "")
                 rpar = Leaf(token.RPAR, "")
@@ -2434,6 +2460,12 @@ def is_vararg(leaf: Leaf, within: Set[NodeType]) -> bool:
     return p.type in within
 
 
+def is_multiline_string(leaf: Leaf) -> bool:
+    """Return True if `leaf` is a multiline string that actually spans many lines."""
+    value = leaf.value.lstrip("furbFURB")
+    return value[:3] in {'"""', "'''"} and "\n" in value
+
+
 def is_stub_suite(node: Node) -> bool:
     """Return True if `node` is a suite with a stub body."""
     if (
@@ -2503,6 +2535,17 @@ def ensure_visible(leaf: Leaf) -> None:
         leaf.value = ")"
 
 
+def should_explode(line: Line, opening_bracket: Leaf) -> bool:
+    """Should `line` immediately be split with `delimiter_split()` after RHS?"""
+    return bool(
+        opening_bracket.parent
+        and opening_bracket.parent.type in {syms.atom, syms.import_from}
+        and opening_bracket.value in "[{("
+        and line.bracket_tracker.delimiters
+        and line.bracket_tracker.max_delimiter_priority() == COMMA_PRIORITY
+    )
+
+
 def is_python36(node: Node) -> bool:
     """Return True if the current file is using Python 3.6+ features.
 
@@ -2631,7 +2674,15 @@ def get_future_imports(node: Node) -> Set[str]:
 
 PYTHON_EXTENSIONS = {".py", ".pyi"}
 BLACKLISTED_DIRECTORIES = {
-    "build", "buck-out", "dist", "_build", ".git", ".hg", ".mypy_cache", ".tox", ".venv"
+    "build",
+    "buck-out",
+    "dist",
+    "_build",
+    ".git",
+    ".hg",
+    ".mypy_cache",
+    ".tox",
+    ".venv",
 }
 
 
@@ -2833,7 +2884,7 @@ def diff(a: str, b: str, a_name: str, b_name: str) -> str:
     )
 
 
-def cancel(tasks: List[asyncio.Task]) -> None:
+def cancel(tasks: Iterable[asyncio.Task]) -> None:
     """asyncio signal handler that cancels all `tasks` and reports to stderr."""
     err("Aborted!")
     for task in tasks: