]> 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:

fix type errors in setup.py (#179)
[etc/vim.git] / black.py
index 9f230e794c57f8e45781fa0566495f91a897d812..4c5f0f0c2590d476e53fdb3f5d19ed3152928a06 100644 (file)
--- a/black.py
+++ b/black.py
@@ -43,8 +43,9 @@ from blib2to3 import pygram, pytree
 from blib2to3.pgen2 import driver, token
 from blib2to3.pgen2.parse import ParseError
 
-__version__ = "18.4a2"
+__version__ = "18.4a4"
 DEFAULT_LINE_LENGTH = 88
+
 # types
 syms = pygram.python_symbols
 FileContent = str
@@ -184,47 +185,45 @@ def main(
             sources.append(Path("-"))
         else:
             err(f"invalid path: {s}")
-    if check and diff:
-        exc = click.ClickException("Options --check and --diff are mutually exclusive")
-        exc.exit_code = 2
-        raise exc
 
-    if check:
+    if check and not diff:
         write_back = WriteBack.NO
     elif diff:
         write_back = WriteBack.DIFF
     else:
         write_back = WriteBack.YES
+    report = Report(check=check, quiet=quiet)
     if len(sources) == 0:
         ctx.exit(0)
         return
 
     elif len(sources) == 1:
-        return_code = reformat_one(sources[0], line_length, fast, quiet, write_back)
+        reformat_one(sources[0], line_length, fast, write_back, report)
     else:
         loop = asyncio.get_event_loop()
         executor = ProcessPoolExecutor(max_workers=os.cpu_count())
-        return_code = 1
         try:
-            return_code = loop.run_until_complete(
+            loop.run_until_complete(
                 schedule_formatting(
-                    sources, line_length, write_back, fast, quiet, loop, executor
+                    sources, line_length, fast, write_back, report, loop, executor
                 )
             )
         finally:
             shutdown(loop)
-    ctx.exit(return_code)
+        if not quiet:
+            out("All done! ✨ 🍰 ✨")
+            click.echo(str(report))
+    ctx.exit(report.return_code)
 
 
 def reformat_one(
-    src: Path, line_length: int, fast: bool, quiet: bool, write_back: WriteBack
-) -> int:
+    src: Path, line_length: int, fast: bool, write_back: WriteBack, report: "Report"
+) -> None:
     """Reformat a single file under `src` without spawning child processes.
 
     If `quiet` is True, non-error messages are not output. `line_length`,
     `write_back`, and `fast` options are passed to :func:`format_file_in_place`.
     """
-    report = Report(check=write_back is WriteBack.NO, quiet=quiet)
     try:
         changed = Changed.NO
         if not src.is_file() and str(src) == "-":
@@ -235,7 +234,7 @@ def reformat_one(
         else:
             cache: Cache = {}
             if write_back != WriteBack.DIFF:
-                cache = read_cache()
+                cache = read_cache(line_length)
                 src = src.resolve()
                 if src in cache and cache[src] == get_cache_info(src):
                     changed = Changed.CACHED
@@ -246,23 +245,22 @@ def reformat_one(
                 )
             ):
                 changed = Changed.YES
-            if write_back != WriteBack.DIFF and changed is not Changed.NO:
-                write_cache(cache, [src])
+            if write_back == WriteBack.YES and changed is not Changed.NO:
+                write_cache(cache, [src], line_length)
         report.done(src, changed)
     except Exception as exc:
         report.failed(src, str(exc))
-    return report.return_code
 
 
 async def schedule_formatting(
     sources: List[Path],
     line_length: int,
-    write_back: WriteBack,
     fast: bool,
-    quiet: bool,
+    write_back: WriteBack,
+    report: "Report",
     loop: BaseEventLoop,
     executor: Executor,
-) -> int:
+) -> None:
     """Run formatting of `sources` in parallel using the provided `executor`.
 
     (Use ProcessPoolExecutors for actual parallelism.)
@@ -270,10 +268,9 @@ async def schedule_formatting(
     `line_length`, `write_back`, and `fast` options are passed to
     :func:`format_file_in_place`.
     """
-    report = Report(check=write_back is WriteBack.NO, quiet=quiet)
     cache: Cache = {}
     if write_back != WriteBack.DIFF:
-        cache = read_cache()
+        cache = read_cache(line_length)
         sources, cached = filter_cached(cache, sources)
         for src in cached:
             report.done(src, Changed.CACHED)
@@ -293,8 +290,12 @@ async def schedule_formatting(
             for src in sources
         }
         _task_values = list(tasks.values())
-        loop.add_signal_handler(signal.SIGINT, cancel, _task_values)
-        loop.add_signal_handler(signal.SIGTERM, cancel, _task_values)
+        try:
+            loop.add_signal_handler(signal.SIGINT, cancel, _task_values)
+            loop.add_signal_handler(signal.SIGTERM, cancel, _task_values)
+        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():
@@ -311,15 +312,8 @@ async def schedule_formatting(
 
     if cancelled:
         await asyncio.gather(*cancelled, loop=loop, return_exceptions=True)
-    elif not quiet:
-        out("All done! ✨ 🍰 ✨")
-    if not quiet:
-        click.echo(str(report))
-
-    if write_back != WriteBack.DIFF and formatted:
-        write_cache(cache, formatted)
-
-    return report.return_code
+    if write_back == WriteBack.YES and formatted:
+        write_cache(cache, formatted, line_length)
 
 
 def format_file_in_place(
@@ -348,8 +342,8 @@ def format_file_in_place(
         with open(src, "w", encoding=src_buffer.encoding) as f:
             f.write(dst_contents)
     elif write_back == write_back.DIFF:
-        src_name = f"{src.name}  (original)"
-        dst_name = f"{src.name}  (formatted)"
+        src_name = f"{src}  (original)"
+        dst_name = f"{src}  (formatted)"
         diff_contents = diff(src_contents, dst_contents, src_name, dst_name)
         if lock:
             lock.acquire()
@@ -435,7 +429,6 @@ def format_str(src_contents: str, line_length: int) -> FileContent:
 GRAMMARS = [
     pygram.python_grammar_no_print_statement_no_exec_statement,
     pygram.python_grammar_no_print_statement,
-    pygram.python_grammar_no_exec_statement,
     pygram.python_grammar,
 ]
 
@@ -591,6 +584,7 @@ UNPACKING_PARENTS = {
 }
 COMPREHENSION_PRIORITY = 20
 COMMA_PRIORITY = 10
+TERNARY_PRIORITY = 7
 LOGIC_PRIORITY = 5
 STRING_PRIORITY = 4
 COMPARATOR_PRIORITY = 3
@@ -605,6 +599,8 @@ class BracketTracker:
     bracket_match: Dict[Tuple[Depth, NodeType], Leaf] = Factory(dict)
     delimiters: Dict[LeafID, Priority] = Factory(dict)
     previous: Optional[Leaf] = None
+    _for_loop_variable: bool = False
+    _lambda_arguments: bool = False
 
     def mark(self, leaf: Leaf) -> None:
         """Mark `leaf` with bracket-related metadata. Keep track of delimiters.
@@ -624,6 +620,8 @@ class BracketTracker:
         if leaf.type == token.COMMENT:
             return
 
+        self.maybe_decrement_after_for_loop_variable(leaf)
+        self.maybe_decrement_after_lambda_arguments(leaf)
         if leaf.type in CLOSING_BRACKETS:
             self.depth -= 1
             opening_bracket = self.bracket_match.pop((self.depth, leaf.type))
@@ -641,6 +639,8 @@ class BracketTracker:
             self.bracket_match[self.depth, BRACKET[leaf.type]] = leaf
             self.depth += 1
         self.previous = leaf
+        self.maybe_increment_lambda_arguments(leaf)
+        self.maybe_increment_for_loop_variable(leaf)
 
     def any_open_brackets(self) -> bool:
         """Return True if there is an yet unmatched open bracket on the line."""
@@ -649,11 +649,55 @@ class BracketTracker:
     def max_delimiter_priority(self, exclude: Iterable[LeafID] = ()) -> int:
         """Return the highest priority of a delimiter found on the line.
 
-        Values are consistent with what `is_delimiter()` returns.
+        Values are consistent with what `is_split_*_delimiter()` return.
         Raises ValueError on no delimiters.
         """
         return max(v for k, v in self.delimiters.items() if k not in exclude)
 
+    def maybe_increment_for_loop_variable(self, leaf: Leaf) -> bool:
+        """In a for loop, or comprehension, the variables are often unpacks.
+
+        To avoid splitting on the comma in this situation, increase the depth of
+        tokens between `for` and `in`.
+        """
+        if leaf.type == token.NAME and leaf.value == "for":
+            self.depth += 1
+            self._for_loop_variable = True
+            return True
+
+        return False
+
+    def maybe_decrement_after_for_loop_variable(self, leaf: Leaf) -> bool:
+        """See `maybe_increment_for_loop_variable` above for explanation."""
+        if self._for_loop_variable and leaf.type == token.NAME and leaf.value == "in":
+            self.depth -= 1
+            self._for_loop_variable = False
+            return True
+
+        return False
+
+    def maybe_increment_lambda_arguments(self, leaf: Leaf) -> bool:
+        """In a lambda expression, there might be more than one argument.
+
+        To avoid splitting on the comma in this situation, increase the depth of
+        tokens between `lambda` and `:`.
+        """
+        if leaf.type == token.NAME and leaf.value == "lambda":
+            self.depth += 1
+            self._lambda_arguments = True
+            return True
+
+        return False
+
+    def maybe_decrement_after_lambda_arguments(self, leaf: Leaf) -> bool:
+        """See `maybe_increment_lambda_arguments` above for explanation."""
+        if self._lambda_arguments and leaf.type == token.COLON:
+            self.depth -= 1
+            self._lambda_arguments = False
+            return True
+
+        return False
+
 
 @dataclass
 class Line:
@@ -664,8 +708,6 @@ class Line:
     comments: List[Tuple[Index, Leaf]] = Factory(list)
     bracket_tracker: BracketTracker = Factory(BracketTracker)
     inside_brackets: bool = False
-    has_for: bool = False
-    _for_loop_variable: bool = False
 
     def append(self, leaf: Leaf, preformatted: bool = False) -> None:
         """Add a new `leaf` to the end of the line.
@@ -686,10 +728,8 @@ class Line:
             # imports, for which we only preserve newlines.
             leaf.prefix += whitespace(leaf)
         if self.inside_brackets or not preformatted:
-            self.maybe_decrement_after_for_loop_variable(leaf)
             self.bracket_tracker.mark(leaf)
             self.maybe_remove_trailing_comma(leaf)
-            self.maybe_increment_for_loop_variable(leaf)
 
         if not self.append_comment(leaf):
             self.leaves.append(leaf)
@@ -836,29 +876,6 @@ class Line:
 
         return False
 
-    def maybe_increment_for_loop_variable(self, leaf: Leaf) -> bool:
-        """In a for loop, or comprehension, the variables are often unpacks.
-
-        To avoid splitting on the comma in this situation, increase the depth of
-        tokens between `for` and `in`.
-        """
-        if leaf.type == token.NAME and leaf.value == "for":
-            self.has_for = True
-            self.bracket_tracker.depth += 1
-            self._for_loop_variable = True
-            return True
-
-        return False
-
-    def maybe_decrement_after_for_loop_variable(self, leaf: Leaf) -> bool:
-        """See `maybe_increment_for_loop_variable` above for explanation."""
-        if self._for_loop_variable and leaf.type == token.NAME and leaf.value == "in":
-            self.bracket_tracker.depth -= 1
-            self._for_loop_variable = False
-            return True
-
-        return False
-
     def append_comment(self, comment: Leaf) -> bool:
         """Add an inline or standalone comment to the line."""
         if (
@@ -1025,8 +1042,14 @@ class EmptyLineTracker:
                 # Don't insert empty lines before the first line in the file.
                 return 0, 0
 
-            if self.previous_line and self.previous_line.is_decorator:
-                # Don't insert empty lines between decorators.
+            if self.previous_line.is_decorator:
+                return 0, 0
+
+            if (
+                self.previous_line.is_comment
+                and self.previous_line.depth == current_line.depth
+                and before == 0
+            ):
                 return 0, 0
 
             newlines = 2
@@ -1034,9 +1057,6 @@ class EmptyLineTracker:
                 newlines -= 1
             return newlines, 0
 
-        if current_line.is_flow_control:
-            return before, 1
-
         if (
             self.previous_line
             and self.previous_line.is_import
@@ -1045,13 +1065,6 @@ class EmptyLineTracker:
         ):
             return (before or 1), 0
 
-        if (
-            self.previous_line
-            and self.previous_line.is_yield
-            and (not current_line.is_yield or depth != self.previous_line.depth)
-        ):
-            return (before or 1), 0
-
         return before, 0
 
 
@@ -1143,7 +1156,16 @@ class LineGenerator(Visitor[Line]):
 
     def visit_DEDENT(self, node: Node) -> Iterator[Line]:
         """Decrease indentation level, maybe yield a line."""
-        # DEDENT has no value. Additionally, in blib2to3 it never holds comments.
+        # The current line might still wait for trailing comments.  At DEDENT time
+        # there won't be any (they would be prefixes on the preceding NEWLINE).
+        # Emit the line then.
+        yield from self.line()
+
+        # While DEDENT has no value, its prefix may contain standalone comments
+        # that belong to the current indentation level.  Get 'em.
+        yield from self.visit_default(node)
+
+        # Finally, emit the dedent.
         yield from self.line(-1)
 
     def visit_stmt(
@@ -1352,17 +1374,11 @@ def whitespace(leaf: Leaf) -> str:  # noqa C901
 
     if p.type in {syms.parameters, syms.arglist}:
         # untyped function signatures or calls
-        if t == token.RPAR:
-            return NO
-
         if not prev or prev.type != token.COMMA:
             return NO
 
     elif p.type == syms.varargslist:
         # lambdas
-        if t == token.RPAR:
-            return NO
-
         if prev and prev.type != token.COMMA:
             return NO
 
@@ -1456,21 +1472,9 @@ def whitespace(leaf: Leaf) -> str:  # noqa C901
             # dots, but not the first one.
             return NO
 
-    elif (
-        p.type == syms.listmaker
-        or p.type == syms.testlist_gexp
-        or p.type == syms.subscriptlist
-    ):
-        # list interior, including unpacking
-        if not prev:
-            return NO
-
     elif p.type == syms.dictsetmaker:
-        # dict and set interior, including unpacking
-        if not prev:
-            return NO
-
-        if prev.type == token.DOUBLESTAR:
+        # dict unpacking
+        if prev and prev.type == token.DOUBLESTAR:
             return NO
 
     elif p.type in {syms.factor, syms.star_expr}:
@@ -1590,23 +1594,20 @@ def is_split_before_delimiter(leaf: Leaf, previous: Leaf = None) -> int:
     ):
         return COMPREHENSION_PRIORITY
 
+    if (
+        leaf.type == token.NAME
+        and leaf.value in {"if", "else"}
+        and leaf.parent
+        and leaf.parent.type == syms.test
+    ):
+        return TERNARY_PRIORITY
+
     if leaf.type == token.NAME and leaf.value in LOGIC_OPERATORS and leaf.parent:
         return LOGIC_PRIORITY
 
     return 0
 
 
-def is_delimiter(leaf: Leaf, previous: Leaf = None) -> int:
-    """Return the priority of the `leaf` delimiter. Return 0 if not delimiter.
-
-    Higher numbers are higher priority.
-    """
-    return max(
-        is_split_before_delimiter(leaf, previous),
-        is_split_after_delimiter(leaf, previous),
-    )
-
-
 def generate_comments(leaf: Leaf) -> Iterator[Leaf]:
     """Clean the prefix of the `leaf` and generate comments from it, if any.
 
@@ -1712,6 +1713,8 @@ def split_line(
     split_funcs: List[SplitFunc]
     if line.is_def:
         split_funcs = [left_hand_split]
+    elif line.is_import:
+        split_funcs = [explode_split]
     elif line.inside_brackets:
         split_funcs = [delimiter_split, standalone_comment_split, right_hand_split]
     else:
@@ -1978,6 +1981,25 @@ 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
@@ -2165,6 +2187,11 @@ def is_vararg(leaf: Leaf, within: Set[NodeType]) -> bool:
 
 
 def max_delimiter_priority_in_atom(node: LN) -> int:
+    """Return maximum delimiter priority inside `node`.
+
+    This is specific to atoms with contents contained in a pair of parentheses.
+    If `node` isn't an atom or there are no enclosing parentheses, returns 0.
+    """
     if node.type != syms.atom:
         return 0
 
@@ -2467,18 +2494,22 @@ def sub_twice(regex: Pattern[str], replacement: str, original: str) -> str:
 
 
 CACHE_DIR = Path(user_cache_dir("black", version=__version__))
-CACHE_FILE = CACHE_DIR / "cache.pickle"
 
 
-def read_cache() -> Cache:
+def get_cache_file(line_length: int) -> Path:
+    return CACHE_DIR / f"cache.{line_length}.pickle"
+
+
+def read_cache(line_length: int) -> Cache:
     """Read the cache if it exists and is well formed.
 
     If it is not well formed, the call to write_cache later should resolve the issue.
     """
-    if not CACHE_FILE.exists():
+    cache_file = get_cache_file(line_length)
+    if not cache_file.exists():
         return {}
 
-    with CACHE_FILE.open("rb") as fobj:
+    with cache_file.open("rb") as fobj:
         try:
             cache: Cache = pickle.load(fobj)
         except pickle.UnpicklingError:
@@ -2511,13 +2542,14 @@ def filter_cached(
     return todo, done
 
 
-def write_cache(cache: Cache, sources: List[Path]) -> None:
+def write_cache(cache: Cache, sources: List[Path], line_length: int) -> None:
     """Update the cache file."""
+    cache_file = get_cache_file(line_length)
     try:
         if not CACHE_DIR.exists():
             CACHE_DIR.mkdir(parents=True)
         new_cache = {**cache, **{src.resolve(): get_cache_info(src) for src in sources}}
-        with CACHE_FILE.open("wb") as fobj:
+        with cache_file.open("wb") as fobj:
             pickle.dump(new_cache, fobj, protocol=pickle.HIGHEST_PROTOCOL)
     except OSError:
         pass