]> 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:

Preserve line endings when formatting a file in place (#288)
[etc/vim.git] / black.py
index ce4a44f392f3daadf1e1f4be88aded372a6c7d96..36e49b026a53492ba635621353c2e37ffa2b9520 100644 (file)
--- a/black.py
+++ b/black.py
@@ -4,6 +4,7 @@ from asyncio.base_events import BaseEventLoop
 from concurrent.futures import Executor, ProcessPoolExecutor
 from enum import Enum, Flag
 from functools import partial, wraps
+import io
 import keyword
 import logging
 from multiprocessing import Manager
@@ -47,7 +48,7 @@ from blib2to3.pgen2.parse import ParseError
 __version__ = "18.5b1"
 DEFAULT_LINE_LENGTH = 88
 DEFAULT_EXCLUDES = (
-    r"build/|buck-out/|dist/|_build/|\.git/|\.hg/|\.mypy_cache/|\.tox/|\.venv/"
+    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__))
@@ -119,6 +120,13 @@ class WriteBack(Enum):
     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
@@ -132,6 +140,19 @@ class FileMode(Flag):
     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(
@@ -143,55 +164,46 @@ class FileMode(Flag):
     show_default=True,
 )
 @click.option(
-    "--check",
+    "--py36",
     is_flag=True,
     help=(
-        "Don't write the files back, just return the status.  Return code 0 "
-        "means nothing would change.  Return code 1 means some files would be "
-        "reformatted.  Return code 123 means there was an internal error."
+        "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(
-    "--diff",
-    is_flag=True,
-    help="Don't write the files back, just output a diff for each file on stdout.",
-)
-@click.option(
-    "--fast/--safe",
+    "--pyi",
     is_flag=True,
-    help="If --fast given, skip temporary sanity checks. [default: --safe]",
+    help=(
+        "Format all input files like typing stubs regardless of file extension "
+        "(useful when piping source on standard input)."
+    ),
 )
 @click.option(
-    "-q",
-    "--quiet",
+    "-S",
+    "--skip-string-normalization",
     is_flag=True,
-    help=(
-        "Don't emit non-error messages to stderr. Errors are still emitted, "
-        "silence those with 2>/dev/null."
-    ),
+    help="Don't normalize string quotes or prefixes.",
 )
 @click.option(
-    "--pyi",
+    "--check",
     is_flag=True,
     help=(
-        "Consider all input files typing stubs regardless of file extension "
-        "(useful when piping source on standard input)."
+        "Don't write the files back, just return the status.  Return code 0 "
+        "means nothing would change.  Return code 1 means some files would be "
+        "reformatted.  Return code 123 means there was an internal error."
     ),
 )
 @click.option(
-    "--py36",
+    "--diff",
     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]"
-    ),
+    help="Don't write the files back, just output a diff for each file on stdout.",
 )
 @click.option(
-    "-S",
-    "--skip-string-normalization",
+    "--fast/--safe",
     is_flag=True,
-    help="Don't normalize string quotes or prefixes.",
+    help="If --fast given, skip temporary sanity checks. [default: --safe]",
 )
 @click.option(
     "--include",
@@ -199,8 +211,10 @@ class FileMode(Flag):
     default=DEFAULT_INCLUDES,
     help=(
         "A regular expression that matches files and directories that should be "
-        "included on recursive searches. On Windows, use forward slashes for "
-        "directories."
+        "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,
 )
@@ -210,11 +224,30 @@ class FileMode(Flag):
     default=DEFAULT_EXCLUDES,
     help=(
         "A regular expression that matches files and directories that should be "
-        "excluded on recursive searches. On Windows, use forward slashes for "
-        "directories."
+        "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",
+    is_flag=True,
+    help=(
+        "Don't emit non-error messages to stderr. Errors are still emitted, "
+        "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",
@@ -234,11 +267,17 @@ def main(
     py36: bool,
     skip_string_normalization: bool,
     quiet: bool,
+    verbose: bool,
     include: str,
     exclude: str,
     src: List[str],
 ) -> None:
     """The uncompromising code formatter."""
+    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: List[Path] = []
     try:
         include_regex = re.compile(include)
@@ -250,32 +289,18 @@ def main(
     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():
-            sources.extend(gen_python_files_in_dir(p, include_regex, exclude_regex))
-        elif p.is_file():
+            sources.extend(
+                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
             sources.append(p)
-        elif s == "-":
-            sources.append(Path("-"))
         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
-    mode = FileMode.AUTO_DETECT
-    if py36:
-        mode |= FileMode.PYTHON36
-    if pyi:
-        mode |= FileMode.PYI
-    if skip_string_normalization:
-        mode |= FileMode.NO_STRING_NORMALIZATION
-    report = Report(check=check, quiet=quiet)
     if len(sources) == 0:
         out("No paths given. Nothing to do 😴")
         ctx.exit(0)
@@ -339,8 +364,8 @@ def reformat_one(
             cache: Cache = {}
             if write_back != WriteBack.DIFF:
                 cache = read_cache(line_length, mode)
-                src = src.resolve()
-                if src in cache and cache[src] == get_cache_info(src):
+                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(
                 src,
@@ -441,8 +466,9 @@ def format_file_in_place(
     """
     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(
             src_contents, line_length=line_length, fast=fast, mode=mode
@@ -451,7 +477,7 @@ def format_file_in_place(
         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)"
@@ -460,7 +486,14 @@ def format_file_in_place(
         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()
@@ -479,7 +512,7 @@ def format_stdin_to_stdout(
     `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 = format_file_contents(src, line_length=line_length, fast=fast, mode=mode)
@@ -490,11 +523,25 @@ def format_stdin_to_stdout(
 
     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)"
-            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(
@@ -555,6 +602,19 @@ def format_str(
     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,
@@ -566,8 +626,7 @@ 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":
-        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:
@@ -2789,42 +2848,72 @@ def get_future_imports(node: Node) -> Set[str]:
 
 
 def gen_python_files_in_dir(
-    path: Path, include: Pattern[str], exclude: Pattern[str]
+    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.
-    """
 
+    `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():
-        searchable_path = str(child.as_posix())
-        if Path(child.parts[0]).is_dir():
-            searchable_path = "/" + searchable_path
+        normalized_path = "/" + child.resolve().relative_to(root).as_posix()
         if child.is_dir():
-            searchable_path = searchable_path + "/"
-            exclude_match = exclude.search(searchable_path)
-            if exclude_match and len(exclude_match.group()) > 0:
-                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
 
-            yield from gen_python_files_in_dir(child, include, exclude)
+        if child.is_dir():
+            yield from gen_python_files_in_dir(child, root, include, exclude, report)
 
-        else:
-            include_match = include.search(searchable_path)
-            exclude_match = exclude.search(searchable_path)
-            if (
-                child.is_file()
-                and include_match
-                and len(include_match.group()) > 0
-                and (not exclude_match or len(exclude_match.group()) == 0)
-            ):
+        elif child.is_file():
+            include_match = include.search(normalized_path)
+            if include_match:
                 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)`."""
 
     check: bool = False
     quiet: bool = False
+    verbose: bool = False
     change_count: int = 0
     same_count: int = 0
     failure_count: int = 0
@@ -2833,11 +2922,11 @@ 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"
-            if not self.quiet:
+            if self.verbose or not self.quiet:
                 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:
@@ -2850,6 +2939,10 @@ class Report:
         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.