Always show summary of reformatting

[etc/vim.git] / black.py

diff --git a/black.py b/black.py
index afc37d99fe3ac547383c1fa3c59cfba763aefc51..19a023cff52db2e92d5cc0950125e20c96d24dff 100644 (file)
--- a/black.py
+++ b/black.py
@@ -2,8 +2,9 @@ import asyncio
 import pickle
 from asyncio.base_events import BaseEventLoop
 from concurrent.futures import Executor, ProcessPoolExecutor
 import pickle
 from asyncio.base_events import BaseEventLoop
 from concurrent.futures import Executor, ProcessPoolExecutor

-from enum import Enum
+from enum import Enum, Flag

 from functools import partial, wraps
 from functools import partial, wraps

+import io

 import keyword
 import logging
 from multiprocessing import Manager
 import keyword
 import logging
 from multiprocessing import Manager


@@ -44,8 +45,12 @@ from blib2to3.pgen2 import driver, token
 from blib2to3.pgen2.parse import ParseError
 
 
 from blib2to3.pgen2.parse import ParseError
 
 

-__version__ = "18.5b0"
+__version__ = "18.5b1"

 DEFAULT_LINE_LENGTH = 88
 DEFAULT_LINE_LENGTH = 88

+DEFAULT_EXCLUDES = (
+    r"/(\.git|\.hg|\.mypy_cache|\.tox|\.venv|_build|buck-out|build|dist)/"
+)
+DEFAULT_INCLUDES = r"\.pyi?$"

 CACHE_DIR = Path(user_cache_dir("black", version=__version__))
 
 
 CACHE_DIR = Path(user_cache_dir("black", version=__version__))
 
 


@@ -115,6 +120,13 @@ class WriteBack(Enum):
     YES = 1
     DIFF = 2
 
     YES = 1
     DIFF = 2
 

+    @classmethod
+    def from_configuration(cls, *, check: bool, diff: bool) -> "WriteBack":
+        if check and not diff:
+            return cls.NO
+
+        return cls.DIFF if diff else cls.YES
+

 
 class Changed(Enum):
     NO = 0
 
 class Changed(Enum):
     NO = 0


@@ -122,6 +134,26 @@ class Changed(Enum):
     YES = 2
 
 
     YES = 2
 
 

+class FileMode(Flag):
+    AUTO_DETECT = 0
+    PYTHON36 = 1
+    PYI = 2
+    NO_STRING_NORMALIZATION = 4
+
+    @classmethod
+    def from_configuration(
+        cls, *, py36: bool, pyi: bool, skip_string_normalization: bool
+    ) -> "FileMode":
+        mode = cls.AUTO_DETECT
+        if py36:
+            mode |= cls.PYTHON36
+        if pyi:
+            mode |= cls.PYI
+        if skip_string_normalization:
+            mode |= cls.NO_STRING_NORMALIZATION
+        return mode
+
+

 @click.command()
 @click.option(
     "-l",
 @click.command()
 @click.option(
     "-l",


@@ -131,6 +163,29 @@ class Changed(Enum):
     help="How many character per line to allow.",
     show_default=True,
 )
     help="How many character per line to allow.",
     show_default=True,
 )

+@click.option(
+    "--py36",
+    is_flag=True,
+    help=(
+        "Allow using Python 3.6-only syntax on all input files.  This will put "
+        "trailing commas in function signatures and calls also after *args and "
+        "**kwargs.  [default: per-file auto-detection]"
+    ),
+)
+@click.option(
+    "--pyi",
+    is_flag=True,
+    help=(
+        "Format all input files like typing stubs regardless of file extension "
+        "(useful when piping source on standard input)."
+    ),
+)
+@click.option(
+    "-S",
+    "--skip-string-normalization",
+    is_flag=True,
+    help="Don't normalize string quotes or prefixes.",
+)

 @click.option(
     "--check",
     is_flag=True,
 @click.option(
     "--check",
     is_flag=True,


@@ -150,6 +205,31 @@ class Changed(Enum):
     is_flag=True,
     help="If --fast given, skip temporary sanity checks. [default: --safe]",
 )
     is_flag=True,
     help="If --fast given, skip temporary sanity checks. [default: --safe]",
 )

+@click.option(
+    "--include",
+    type=str,
+    default=DEFAULT_INCLUDES,
+    help=(
+        "A regular expression that matches files and directories that should be "
+        "included on recursive searches.  An empty value means all files are "
+        "included regardless of the name.  Use forward slashes for directories on "
+        "all platforms (Windows, too).  Exclusions are calculated first, inclusions "
+        "later."
+    ),
+    show_default=True,
+)
+@click.option(
+    "--exclude",
+    type=str,
+    default=DEFAULT_EXCLUDES,
+    help=(
+        "A regular expression that matches files and directories that should be "
+        "excluded on recursive searches.  An empty value means no paths are excluded. "
+        "Use forward slashes for directories on all platforms (Windows, too).  "
+        "Exclusions are calculated first, inclusions later."
+    ),
+    show_default=True,
+)

 @click.option(
     "-q",
     "--quiet",
 @click.option(
     "-q",
     "--quiet",


@@ -159,6 +239,15 @@ class Changed(Enum):
         "silence those with 2>/dev/null."
     ),
 )
         "silence those with 2>/dev/null."
     ),
 )

+@click.option(
+    "-v",
+    "--verbose",
+    is_flag=True,
+    help=(
+        "Also emit messages to stderr about files that were not changed or were "
+        "ignored due to --exclude=."
+    ),
+)

 @click.version_option(version=__version__)
 @click.argument(
     "src",
 @click.version_option(version=__version__)
 @click.argument(
     "src",


@@ -174,92 +263,132 @@ def main(
     check: bool,
     diff: bool,
     fast: bool,
     check: bool,
     diff: bool,
     fast: bool,

+    pyi: bool,
+    py36: bool,
+    skip_string_normalization: bool,

     quiet: bool,
     quiet: bool,

+    verbose: bool,
+    include: str,
+    exclude: str,

     src: List[str],
 ) -> None:
     """The uncompromising code formatter."""
     src: List[str],
 ) -> None:
     """The uncompromising code formatter."""

-    sources: List[Path] = []
+    write_back = WriteBack.from_configuration(check=check, diff=diff)
+    mode = FileMode.from_configuration(
+        py36=py36, pyi=pyi, skip_string_normalization=skip_string_normalization
+    )
+    report = Report(check=check, quiet=quiet, verbose=verbose)
+    sources: Set[Path] = set()
+    try:
+        include_regex = re.compile(include)
+    except re.error:
+        err(f"Invalid regular expression for include given: {include!r}")
+        ctx.exit(2)
+    try:
+        exclude_regex = re.compile(exclude)
+    except re.error:
+        err(f"Invalid regular expression for exclude given: {exclude!r}")
+        ctx.exit(2)
+    root = find_project_root(src)

     for s in src:
         p = Path(s)
         if p.is_dir():
     for s in src:
         p = Path(s)
         if p.is_dir():

-            sources.extend(gen_python_files_in_dir(p))
-        elif p.is_file():
+            sources.update(
+                gen_python_files_in_dir(p, root, include_regex, exclude_regex, report)
+            )
+        elif p.is_file() or s == "-":

             # if a file was explicitly given, we don't care about its extension
             # if a file was explicitly given, we don't care about its extension

-            sources.append(p)
-        elif s == "-":
-            sources.append(Path("-"))
+            sources.add(p)

         else:
             err(f"invalid path: {s}")
         else:
             err(f"invalid path: {s}")

-
-    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:
     if len(sources) == 0:

-        out("No paths given. Nothing to do 😴")
+        if verbose or not quiet:
+            out("No paths given. Nothing to do 😴")

         ctx.exit(0)
         return
 
     elif len(sources) == 1:
         ctx.exit(0)
         return
 
     elif len(sources) == 1:

-        reformat_one(sources[0], line_length, fast, write_back, report)
+        reformat_one(
+            src=sources.pop(),
+            line_length=line_length,
+            fast=fast,
+            write_back=write_back,
+            mode=mode,
+            report=report,
+        )

     else:
         loop = asyncio.get_event_loop()
         executor = ProcessPoolExecutor(max_workers=os.cpu_count())
         try:
             loop.run_until_complete(
                 schedule_formatting(
     else:
         loop = asyncio.get_event_loop()
         executor = ProcessPoolExecutor(max_workers=os.cpu_count())
         try:
             loop.run_until_complete(
                 schedule_formatting(

-                    sources, line_length, fast, write_back, report, loop, executor
+                    sources=sources,
+                    line_length=line_length,
+                    fast=fast,
+                    write_back=write_back,
+                    mode=mode,
+                    report=report,
+                    loop=loop,
+                    executor=executor,

                 )
             )
         finally:
             shutdown(loop)
                 )
             )
         finally:
             shutdown(loop)

-        if not quiet:
-            out("All done! ✨ 🍰 ✨")
-            click.echo(str(report))
+    if verbose or not quiet:
+        out("All done! ✨ 🍰 ✨")
+        click.echo(str(report))

     ctx.exit(report.return_code)
 
 
 def reformat_one(
     ctx.exit(report.return_code)
 
 
 def reformat_one(

-    src: Path, line_length: int, fast: bool, write_back: WriteBack, report: "Report"
+    src: Path,
+    line_length: int,
+    fast: bool,
+    write_back: WriteBack,
+    mode: FileMode,
+    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`,
 ) -> 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`.
+    `write_back`, `fast` and `pyi` options are passed to
+    :func:`format_file_in_place` or :func:`format_stdin_to_stdout`.

     """
     try:
         changed = Changed.NO
         if not src.is_file() and str(src) == "-":
             if format_stdin_to_stdout(
     """
     try:
         changed = Changed.NO
         if not src.is_file() and str(src) == "-":
             if format_stdin_to_stdout(

-                line_length=line_length, fast=fast, write_back=write_back
+                line_length=line_length, fast=fast, write_back=write_back, mode=mode

             ):
                 changed = Changed.YES
         else:
             cache: Cache = {}
             if write_back != WriteBack.DIFF:
             ):
                 changed = Changed.YES
         else:
             cache: Cache = {}
             if write_back != WriteBack.DIFF:

-                cache = read_cache(line_length)
-                src = src.resolve()
-                if src in cache and cache[src] == get_cache_info(src):
+                cache = read_cache(line_length, mode)
+                res_src = src.resolve()
+                if res_src in cache and cache[res_src] == get_cache_info(res_src):

                     changed = Changed.CACHED
             if changed is not Changed.CACHED and format_file_in_place(
                     changed = Changed.CACHED
             if changed is not Changed.CACHED and format_file_in_place(

-                src, line_length=line_length, fast=fast, write_back=write_back
+                src,
+                line_length=line_length,
+                fast=fast,
+                write_back=write_back,
+                mode=mode,

             ):
                 changed = Changed.YES
             if write_back == WriteBack.YES and changed is not Changed.NO:
             ):
                 changed = Changed.YES
             if write_back == WriteBack.YES and changed is not Changed.NO:

-                write_cache(cache, [src], line_length)
+                write_cache(cache, [src], line_length, mode)

         report.done(src, changed)
     except Exception as exc:
         report.failed(src, str(exc))
 
 
 async def schedule_formatting(
         report.done(src, changed)
     except Exception as exc:
         report.failed(src, str(exc))
 
 
 async def schedule_formatting(

-    sources: List[Path],
+    sources: Set[Path],

     line_length: int,
     fast: bool,
     write_back: WriteBack,
     line_length: int,
     fast: bool,
     write_back: WriteBack,

+    mode: FileMode,

     report: "Report",
     loop: BaseEventLoop,
     executor: Executor,
     report: "Report",
     loop: BaseEventLoop,
     executor: Executor,


@@ -268,14 +397,14 @@ async def schedule_formatting(
 
     (Use ProcessPoolExecutors for actual parallelism.)
 
 
     (Use ProcessPoolExecutors for actual parallelism.)
 

-    `line_length`, `write_back`, and `fast` options are passed to
+    `line_length`, `write_back`, `fast`, and `pyi` options are passed to

     :func:`format_file_in_place`.
     """
     cache: Cache = {}
     if write_back != WriteBack.DIFF:
     :func:`format_file_in_place`.
     """
     cache: Cache = {}
     if write_back != WriteBack.DIFF:

-        cache = read_cache(line_length)
+        cache = read_cache(line_length, mode)

         sources, cached = filter_cached(cache, sources)
         sources, cached = filter_cached(cache, sources)

-        for src in cached:
+        for src in sorted(cached):

             report.done(src, Changed.CACHED)
     cancelled = []
     formatted = []
             report.done(src, Changed.CACHED)
     cancelled = []
     formatted = []


@@ -288,7 +417,14 @@ async def schedule_formatting(
             lock = manager.Lock()
         tasks = {
             loop.run_in_executor(
             lock = manager.Lock()
         tasks = {
             loop.run_in_executor(

-                executor, format_file_in_place, src, line_length, fast, write_back, lock
+                executor,
+                format_file_in_place,
+                src,
+                line_length,
+                fast,
+                write_back,
+                mode,
+                lock,

             ): src
             for src in sorted(sources)
         }
             ): src
             for src in sorted(sources)
         }


@@ -313,7 +449,7 @@ async def schedule_formatting(
     if cancelled:
         await asyncio.gather(*cancelled, loop=loop, return_exceptions=True)
     if write_back == WriteBack.YES and formatted:
     if cancelled:
         await asyncio.gather(*cancelled, loop=loop, return_exceptions=True)
     if write_back == WriteBack.YES and formatted:

-        write_cache(cache, formatted, line_length)
+        write_cache(cache, formatted, line_length, mode)

 
 
 def format_file_in_place(
 
 
 def format_file_in_place(


@@ -321,6 +457,7 @@ def format_file_in_place(
     line_length: int,
     fast: bool,
     write_back: WriteBack = WriteBack.NO,
     line_length: int,
     fast: bool,
     write_back: WriteBack = WriteBack.NO,

+    mode: FileMode = FileMode.AUTO_DETECT,

     lock: Any = None,  # multiprocessing.Manager().Lock() is some crazy proxy
 ) -> bool:
     """Format file under `src` path. Return True if changed.
     lock: Any = None,  # multiprocessing.Manager().Lock() is some crazy proxy
 ) -> bool:
     """Format file under `src` path. Return True if changed.


@@ -328,19 +465,20 @@ def format_file_in_place(
     If `write_back` is True, write reformatted code back to stdout.
     `line_length` and `fast` options are passed to :func:`format_file_contents`.
     """
     If `write_back` is True, write reformatted code back to stdout.
     `line_length` and `fast` options are passed to :func:`format_file_contents`.
     """

-    is_pyi = src.suffix == ".pyi"
+    if src.suffix == ".pyi":
+        mode |= FileMode.PYI

 
 

-    with tokenize.open(src) as src_buffer:
-        src_contents = src_buffer.read()
+    with open(src, "rb") as buf:
+        newline, encoding, src_contents = prepare_input(buf.read())

     try:
         dst_contents = format_file_contents(
     try:
         dst_contents = format_file_contents(

-            src_contents, line_length=line_length, fast=fast, is_pyi=is_pyi
+            src_contents, line_length=line_length, fast=fast, mode=mode

         )
     except NothingChanged:
         return False
 
     if write_back == write_back.YES:
         )
     except NothingChanged:
         return False
 
     if write_back == write_back.YES:

-        with open(src, "w", encoding=src_buffer.encoding) as f:
+        with open(src, "w", encoding=encoding, newline=newline) as f:

             f.write(dst_contents)
     elif write_back == write_back.DIFF:
         src_name = f"{src}  (original)"
             f.write(dst_contents)
     elif write_back == write_back.DIFF:
         src_name = f"{src}  (original)"


@@ -349,7 +487,14 @@ def format_file_in_place(
         if lock:
             lock.acquire()
         try:
         if lock:
             lock.acquire()
         try:

-            sys.stdout.write(diff_contents)
+            f = io.TextIOWrapper(
+                sys.stdout.buffer,
+                encoding=encoding,
+                newline=newline,
+                write_through=True,
+            )
+            f.write(diff_contents)
+            f.detach()

         finally:
             if lock:
                 lock.release()
         finally:
             if lock:
                 lock.release()


@@ -357,17 +502,21 @@ def format_file_in_place(
 
 
 def format_stdin_to_stdout(
 
 
 def format_stdin_to_stdout(

-    line_length: int, fast: bool, write_back: WriteBack = WriteBack.NO
+    line_length: int,
+    fast: bool,
+    write_back: WriteBack = WriteBack.NO,
+    mode: FileMode = FileMode.AUTO_DETECT,

 ) -> bool:
     """Format file on stdin. Return True if changed.
 
     If `write_back` is True, write reformatted code back to stdout.
 ) -> bool:
     """Format file on stdin. Return True if changed.
 
     If `write_back` is True, write reformatted code back to stdout.

-    `line_length` and `fast` arguments are passed to :func:`format_file_contents`.
+    `line_length`, `fast`, `is_pyi`, and `force_py36` arguments are passed to
+    :func:`format_file_contents`.

     """
     """

-    src = sys.stdin.read()
+    newline, encoding, src = prepare_input(sys.stdin.buffer.read())

     dst = src
     try:
     dst = src
     try:

-        dst = format_file_contents(src, line_length=line_length, fast=fast)
+        dst = format_file_contents(src, line_length=line_length, fast=fast, mode=mode)

         return True
 
     except NothingChanged:
         return True
 
     except NothingChanged:


@@ -375,15 +524,33 @@ def format_stdin_to_stdout(
 
     finally:
         if write_back == WriteBack.YES:
 
     finally:
         if write_back == WriteBack.YES:

-            sys.stdout.write(dst)
+            f = io.TextIOWrapper(
+                sys.stdout.buffer,
+                encoding=encoding,
+                newline=newline,
+                write_through=True,
+            )
+            f.write(dst)
+            f.detach()

         elif write_back == WriteBack.DIFF:
             src_name = "<stdin>  (original)"
             dst_name = "<stdin>  (formatted)"
         elif write_back == WriteBack.DIFF:
             src_name = "<stdin>  (original)"
             dst_name = "<stdin>  (formatted)"

-            sys.stdout.write(diff(src, dst, src_name, dst_name))
+            f = io.TextIOWrapper(
+                sys.stdout.buffer,
+                encoding=encoding,
+                newline=newline,
+                write_through=True,
+            )
+            f.write(diff(src, dst, src_name, dst_name))
+            f.detach()

 
 
 def format_file_contents(
 
 
 def format_file_contents(

-    src_contents: str, *, line_length: int, fast: bool, is_pyi: bool = False
+    src_contents: str,
+    *,
+    line_length: int,
+    fast: bool,
+    mode: FileMode = FileMode.AUTO_DETECT,

 ) -> FileContent:
     """Reformat contents a file and return new contents.
 
 ) -> FileContent:
     """Reformat contents a file and return new contents.
 


@@ -394,20 +561,18 @@ def format_file_contents(
     if src_contents.strip() == "":
         raise NothingChanged
 
     if src_contents.strip() == "":
         raise NothingChanged
 

-    dst_contents = format_str(src_contents, line_length=line_length, is_pyi=is_pyi)
+    dst_contents = format_str(src_contents, line_length=line_length, mode=mode)

     if src_contents == dst_contents:
         raise NothingChanged
 
     if not fast:
         assert_equivalent(src_contents, dst_contents)
     if src_contents == dst_contents:
         raise NothingChanged
 
     if not fast:
         assert_equivalent(src_contents, dst_contents)

-        assert_stable(
-            src_contents, dst_contents, line_length=line_length, is_pyi=is_pyi
-        )
+        assert_stable(src_contents, dst_contents, line_length=line_length, mode=mode)

     return dst_contents
 
 
 def format_str(
     return dst_contents
 
 
 def format_str(

-    src_contents: str, line_length: int, *, is_pyi: bool = False
+    src_contents: str, line_length: int, *, mode: FileMode = FileMode.AUTO_DETECT

 ) -> FileContent:
     """Reformat a string and return new contents.
 
 ) -> FileContent:
     """Reformat a string and return new contents.
 


@@ -416,11 +581,15 @@ def format_str(
     src_node = lib2to3_parse(src_contents)
     dst_contents = ""
     future_imports = get_future_imports(src_node)
     src_node = lib2to3_parse(src_contents)
     dst_contents = ""
     future_imports = get_future_imports(src_node)

-    elt = EmptyLineTracker(is_pyi=is_pyi)
-    py36 = is_python36(src_node)
+    is_pyi = bool(mode & FileMode.PYI)
+    py36 = bool(mode & FileMode.PYTHON36) or is_python36(src_node)
+    normalize_strings = not bool(mode & FileMode.NO_STRING_NORMALIZATION)

     lines = LineGenerator(
     lines = LineGenerator(

-        remove_u_prefix=py36 or "unicode_literals" in future_imports, is_pyi=is_pyi
+        remove_u_prefix=py36 or "unicode_literals" in future_imports,
+        is_pyi=is_pyi,
+        normalize_strings=normalize_strings,

     )
     )

+    elt = EmptyLineTracker(is_pyi=is_pyi)

     empty_line = Line()
     after = 0
     for current_line in lines.visit(src_node):
     empty_line = Line()
     after = 0
     for current_line in lines.visit(src_node):


@@ -434,6 +603,19 @@ def format_str(
     return dst_contents
 
 
     return dst_contents
 
 

+def prepare_input(src: bytes) -> Tuple[str, str, str]:
+    """Analyze `src` and return a tuple of (newline, encoding, decoded_contents)
+
+    Where `newline` is either CRLF or LF, and `decoded_contents` is decoded with
+    universal newlines (i.e. only LF).
+    """
+    srcbuf = io.BytesIO(src)
+    encoding, lines = tokenize.detect_encoding(srcbuf.readline)
+    newline = "\r\n" if b"\r\n" == lines[0][-2:] else "\n"
+    srcbuf.seek(0)
+    return newline, encoding, io.TextIOWrapper(srcbuf, encoding).read()
+
+

 GRAMMARS = [
     pygram.python_grammar_no_print_statement_no_exec_statement,
     pygram.python_grammar_no_print_statement,
 GRAMMARS = [
     pygram.python_grammar_no_print_statement_no_exec_statement,
     pygram.python_grammar_no_print_statement,


@@ -445,8 +627,7 @@ def lib2to3_parse(src_txt: str) -> Node:
     """Given a string with source, return the lib2to3 Node."""
     grammar = pygram.python_grammar_no_print_statement
     if src_txt[-1] != "\n":
     """Given a string with source, return the lib2to3 Node."""
     grammar = pygram.python_grammar_no_print_statement
     if src_txt[-1] != "\n":

-        nl = "\r\n" if "\r\n" in src_txt[:1024] else "\n"
-        src_txt += nl
+        src_txt += "\n"

     for grammar in GRAMMARS:
         drv = driver.Driver(grammar, pytree.convert)
         try:
     for grammar in GRAMMARS:
         drv = driver.Driver(grammar, pytree.convert)
         try:


@@ -878,27 +1059,6 @@ class Line:
             and second_leaf.value == "def"
         )
 
             and second_leaf.value == "def"
         )
 

-    @property
-    def is_flow_control(self) -> bool:
-        """Is this line a flow control statement?
-
-        Those are `return`, `raise`, `break`, and `continue`.
-        """
-        return (
-            bool(self)
-            and self.leaves[0].type == token.NAME
-            and self.leaves[0].value in FLOW_CONTROL
-        )
-
-    @property
-    def is_yield(self) -> bool:
-        """Is this line a yield statement?"""
-        return (
-            bool(self)
-            and self.leaves[0].type == token.NAME
-            and self.leaves[0].value == "yield"
-        )
-

     @property
     def is_class_paren_empty(self) -> bool:
         """Is this a class with no base classes but using parentheses?
     @property
     def is_class_paren_empty(self) -> bool:
         """Is this a class with no base classes but using parentheses?


@@ -915,6 +1075,15 @@ class Line:
             and self.leaves[3].value == ")"
         )
 
             and self.leaves[3].value == ")"
         )
 

+    @property
+    def is_triple_quoted_string(self) -> bool:
+        """Is the line a triple quoted string?"""
+        return (
+            bool(self)
+            and self.leaves[0].type == token.STRING
+            and self.leaves[0].value.startswith(('"""', "'''"))
+        )
+

     def contains_standalone_comments(self, depth_limit: int = sys.maxsize) -> bool:
         """If so, needs to be split before emitting."""
         for leaf in self.leaves:
     def contains_standalone_comments(self, depth_limit: int = sys.maxsize) -> bool:
         """If so, needs to be split before emitting."""
         for leaf in self.leaves:


@@ -1122,6 +1291,7 @@ class EmptyLineTracker:
     the prefix of the first leaf consists of optional newlines.  Those newlines
     are consumed by `maybe_empty_lines()` and included in the computation.
     """
     the prefix of the first leaf consists of optional newlines.  Those newlines
     are consumed by `maybe_empty_lines()` and included in the computation.
     """

+

     is_pyi: bool = False
     previous_line: Optional[Line] = None
     previous_after: int = 0
     is_pyi: bool = False
     previous_line: Optional[Line] = None
     previous_after: int = 0


@@ -1131,8 +1301,7 @@ class EmptyLineTracker:
         """Return the number of extra empty lines before and after the `current_line`.
 
         This is for separating `def`, `async def` and `class` with extra empty
         """Return the number of extra empty lines before and after the `current_line`.
 
         This is for separating `def`, `async def` and `class` with extra empty

-        lines (two on module-level), as well as providing an extra empty line
-        after flow control keywords to make them more prominent.
+        lines (two on module-level).

         """
         if isinstance(current_line, UnformattedLines):
             return 0, 0
         """
         if isinstance(current_line, UnformattedLines):
             return 0, 0


@@ -1173,6 +1342,11 @@ class EmptyLineTracker:
             if self.previous_line.is_decorator:
                 return 0, 0
 
             if self.previous_line.is_decorator:
                 return 0, 0
 

+            if self.previous_line.depth < current_line.depth and (
+                self.previous_line.is_class or self.previous_line.is_def
+            ):
+                return 0, 0
+

             if (
                 self.previous_line.is_comment
                 and self.previous_line.depth == current_line.depth
             if (
                 self.previous_line.is_comment
                 and self.previous_line.depth == current_line.depth


@@ -1204,6 +1378,13 @@ class EmptyLineTracker:
         ):
             return (before or 1), 0
 
         ):
             return (before or 1), 0
 

+        if (
+            self.previous_line
+            and self.previous_line.is_class
+            and current_line.is_triple_quoted_string
+        ):
+            return before, 1
+

         return before, 0
 
 
         return before, 0
 
 


@@ -1214,7 +1395,9 @@ class LineGenerator(Visitor[Line]):
     Note: destroys the tree it's visiting by mutating prefixes of its leaves
     in ways that will no longer stringify to valid Python code on the tree.
     """
     Note: destroys the tree it's visiting by mutating prefixes of its leaves
     in ways that will no longer stringify to valid Python code on the tree.
     """

+

     is_pyi: bool = False
     is_pyi: bool = False

+    normalize_strings: bool = True

     current_line: Line = Factory(Line)
     remove_u_prefix: bool = False
 
     current_line: Line = Factory(Line)
     remove_u_prefix: bool = False
 


@@ -1283,7 +1466,7 @@ class LineGenerator(Visitor[Line]):
 
             else:
                 normalize_prefix(node, inside_brackets=any_open_brackets)
 
             else:
                 normalize_prefix(node, inside_brackets=any_open_brackets)

-                if node.type == token.STRING:
+                if self.normalize_strings and node.type == token.STRING:

                     normalize_string_prefix(node, remove_u_prefix=self.remove_u_prefix)
                     normalize_string_quotes(node)
                 if node.type not in WHITESPACE:
                     normalize_string_prefix(node, remove_u_prefix=self.remove_u_prefix)
                     normalize_string_quotes(node)
                 if node.type not in WHITESPACE:


@@ -2599,7 +2782,8 @@ def generate_trailers_to_omit(line: Line, line_length: int) -> Iterator[Set[Leaf
         if length > line_length:
             break
 
         if length > line_length:
             break
 

-        if leaf.type == STANDALONE_COMMENT:
+        has_inline_comment = leaf_length > len(leaf.value) + len(leaf.prefix)
+        if leaf.type == STANDALONE_COMMENT or has_inline_comment:

             break
 
         optional_brackets.discard(id(leaf))
             break
 
         optional_brackets.discard(id(leaf))


@@ -2664,40 +2848,73 @@ def get_future_imports(node: Node) -> Set[str]:
     return imports
 
 
     return imports
 
 

-PYTHON_EXTENSIONS = {".py", ".pyi"}
-BLACKLISTED_DIRECTORIES = {
-    "build",
-    "buck-out",
-    "dist",
-    "_build",
-    ".git",
-    ".hg",
-    ".mypy_cache",
-    ".tox",
-    ".venv",
-}
-
+def gen_python_files_in_dir(
+    path: Path,
+    root: Path,
+    include: Pattern[str],
+    exclude: Pattern[str],
+    report: "Report",
+) -> Iterator[Path]:
+    """Generate all files under `path` whose paths are not excluded by the
+    `exclude` regex, but are included by the `include` regex.

 
 

-def gen_python_files_in_dir(path: Path) -> Iterator[Path]:
-    """Generate all files under `path` which aren't under BLACKLISTED_DIRECTORIES
-    and have one of the PYTHON_EXTENSIONS.
+    `report` is where output about exclusions goes.

     """
     """

+    assert root.is_absolute(), f"INTERNAL ERROR: `root` must be absolute but is {root}"

     for child in path.iterdir():
     for child in path.iterdir():

+        normalized_path = "/" + child.resolve().relative_to(root).as_posix()

         if child.is_dir():
         if child.is_dir():

-            if child.name in BLACKLISTED_DIRECTORIES:
-                continue
+            normalized_path += "/"
+        exclude_match = exclude.search(normalized_path)
+        if exclude_match and exclude_match.group(0):
+            report.path_ignored(child, f"matches --exclude={exclude.pattern}")
+            continue
+
+        if child.is_dir():
+            yield from gen_python_files_in_dir(child, root, include, exclude, report)
+
+        elif child.is_file():
+            include_match = include.search(normalized_path)
+            if include_match:
+                yield child

 
 

-            yield from gen_python_files_in_dir(child)

 
 

-        elif child.is_file() and child.suffix in PYTHON_EXTENSIONS:
-            yield child
+def find_project_root(srcs: List[str]) -> Path:
+    """Return a directory containing .git, .hg, or pyproject.toml.
+
+    That directory can be one of the directories passed in `srcs` or their
+    common parent.
+
+    If no directory in the tree contains a marker that would specify it's the
+    project root, the root of the file system is returned.
+    """
+    if not srcs:
+        return Path("/").resolve()
+
+    common_base = min(Path(src).resolve() for src in srcs)
+    if common_base.is_dir():
+        # Append a fake file so `parents` below returns `common_base_dir`, too.
+        common_base /= "fake-file"
+    for directory in common_base.parents:
+        if (directory / ".git").is_dir():
+            return directory
+
+        if (directory / ".hg").is_dir():
+            return directory
+
+        if (directory / "pyproject.toml").is_file():
+            return directory
+
+    return directory

 
 
 @dataclass
 class Report:
     """Provides a reformatting counter. Can be rendered with `str(report)`."""
 
 
 @dataclass
 class Report:
     """Provides a reformatting counter. Can be rendered with `str(report)`."""

+

     check: bool = False
     quiet: bool = False
     check: bool = False
     quiet: bool = False

+    verbose: bool = False

     change_count: int = 0
     same_count: int = 0
     failure_count: int = 0
     change_count: int = 0
     same_count: int = 0
     failure_count: int = 0


@@ -2706,11 +2923,11 @@ class Report:
         """Increment the counter for successful reformatting. Write out a message."""
         if changed is Changed.YES:
             reformatted = "would reformat" if self.check else "reformatted"
         """Increment the counter for successful reformatting. Write out a message."""
         if changed is Changed.YES:
             reformatted = "would reformat" if self.check else "reformatted"

-            if not self.quiet:
+            if self.verbose or not self.quiet:

                 out(f"{reformatted} {src}")
             self.change_count += 1
         else:
                 out(f"{reformatted} {src}")
             self.change_count += 1
         else:

-            if not self.quiet:
+            if self.verbose:

                 if changed is Changed.NO:
                     msg = f"{src} already well formatted, good job."
                 else:
                 if changed is Changed.NO:
                     msg = f"{src} already well formatted, good job."
                 else:


@@ -2723,6 +2940,10 @@ class Report:
         err(f"error: cannot format {src}: {message}")
         self.failure_count += 1
 
         err(f"error: cannot format {src}: {message}")
         self.failure_count += 1
 

+    def path_ignored(self, path: Path, message: str) -> None:
+        if self.verbose:
+            out(f"{path} ignored: {message}", bold=False)
+

     @property
     def return_code(self) -> int:
         """Return the exit code that the app should use.
     @property
     def return_code(self) -> int:
         """Return the exit code that the app should use.


@@ -2835,9 +3056,11 @@ def assert_equivalent(src: str, dst: str) -> None:
         ) from None
 
 
         ) from None
 
 

-def assert_stable(src: str, dst: str, line_length: int, is_pyi: bool = False) -> None:
+def assert_stable(
+    src: str, dst: str, line_length: int, mode: FileMode = FileMode.AUTO_DETECT
+) -> None:

     """Raise AssertionError if `dst` reformats differently the second time."""
     """Raise AssertionError if `dst` reformats differently the second time."""

-    newdst = format_str(dst, line_length=line_length, is_pyi=is_pyi)
+    newdst = format_str(dst, line_length=line_length, mode=mode)

     if dst != newdst:
         log = dump_to_file(
             diff(src, dst, "source", "first pass"),
     if dst != newdst:
         log = dump_to_file(
             diff(src, dst, "source", "first pass"),


@@ -2940,9 +3163,6 @@ def enumerate_with_length(
 
         comment: Optional[Leaf]
         for comment in line.comments_after(leaf, index):
 
         comment: Optional[Leaf]
         for comment in line.comments_after(leaf, index):

-            if "\n" in comment.prefix:
-                return  # Oops, standalone comment!
-

             length += len(comment.value)
 
         yield index, leaf, length
             length += len(comment.value)
 
         yield index, leaf, length


@@ -3051,16 +3271,21 @@ def can_omit_invisible_parens(line: Line, line_length: int) -> bool:
     return False
 
 
     return False
 
 

-def get_cache_file(line_length: int) -> Path:
-    return CACHE_DIR / f"cache.{line_length}.pickle"
+def get_cache_file(line_length: int, mode: FileMode) -> Path:
+    pyi = bool(mode & FileMode.PYI)
+    py36 = bool(mode & FileMode.PYTHON36)
+    return (
+        CACHE_DIR
+        / f"cache.{line_length}{'.pyi' if pyi else ''}{'.py36' if py36 else ''}.pickle"
+    )

 
 
 
 

-def read_cache(line_length: int) -> Cache:
+def read_cache(line_length: int, mode: FileMode) -> 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.
     """
     """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.
     """

-    cache_file = get_cache_file(line_length)
+    cache_file = get_cache_file(line_length, mode)

     if not cache_file.exists():
         return {}
 
     if not cache_file.exists():
         return {}
 


@@ -3079,27 +3304,27 @@ def get_cache_info(path: Path) -> CacheInfo:
     return stat.st_mtime, stat.st_size
 
 
     return stat.st_mtime, stat.st_size
 
 

-def filter_cached(
-    cache: Cache, sources: Iterable[Path]
-) -> Tuple[List[Path], List[Path]]:
-    """Split a list of paths into two.
+def filter_cached(cache: Cache, sources: Iterable[Path]) -> Tuple[Set[Path], Set[Path]]:
+    """Split an iterable of paths in `sources` into two sets.

 
 

-    The first list contains paths of files that modified on disk or are not in the
-    cache. The other list contains paths to non-modified files.
+    The first contains paths of files that modified on disk or are not in the
+    cache. The other contains paths to non-modified files.

     """
     """

-    todo, done = [], []
+    todo, done = set(), set()

     for src in sources:
         src = src.resolve()
         if cache.get(src) != get_cache_info(src):
     for src in sources:
         src = src.resolve()
         if cache.get(src) != get_cache_info(src):

-            todo.append(src)
+            todo.add(src)

         else:
         else:

-            done.append(src)
+            done.add(src)

     return todo, done
 
 
     return todo, done
 
 

-def write_cache(cache: Cache, sources: List[Path], line_length: int) -> None:
+def write_cache(
+    cache: Cache, sources: Iterable[Path], line_length: int, mode: FileMode
+) -> None:

     """Update the cache file."""
     """Update the cache file."""

-    cache_file = get_cache_file(line_length)
+    cache_file = get_cache_file(line_length, mode)

     try:
         if not CACHE_DIR.exists():
             CACHE_DIR.mkdir(parents=True)
     try:
         if not CACHE_DIR.exists():
             CACHE_DIR.mkdir(parents=True)