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

Remove blib2to3 grammar cache logging (#3193)
[etc/vim.git] / src / black / __init__.py
index f2efdec83b22a53d1ad00f6d1b1057dcae1f0c0f..2a5c750a5835562ddf5c743b01ab0cc58b4cb75c 100644 (file)
@@ -1,21 +1,22 @@
 import asyncio
-from json.decoder import JSONDecodeError
-import json
-from concurrent.futures import Executor, ThreadPoolExecutor, ProcessPoolExecutor
-from contextlib import contextmanager
-from datetime import datetime
-from enum import Enum
 import io
-from multiprocessing import Manager, freeze_support
+import json
 import os
-from pathlib import Path
-from pathspec.patterns.gitwildmatch import GitWildMatchPatternError
+import platform
 import re
 import signal
 import sys
 import tokenize
 import traceback
+from contextlib import contextmanager
+from dataclasses import replace
+from datetime import datetime
+from enum import Enum
+from json.decoder import JSONDecodeError
+from multiprocessing import Manager, freeze_support
+from pathlib import Path
 from typing import (
+    TYPE_CHECKING,
     Any,
     Dict,
     Generator,
@@ -24,6 +25,7 @@ from typing import (
     MutableMapping,
     Optional,
     Pattern,
+    Sequence,
     Set,
     Sized,
     Tuple,
@@ -31,42 +33,65 @@ from typing import (
 )
 
 import click
-from dataclasses import replace
+from click.core import ParameterSource
 from mypy_extensions import mypyc_attr
+from pathspec.patterns.gitwildmatch import GitWildMatchPatternError
 
-from black.const import DEFAULT_LINE_LENGTH, DEFAULT_INCLUDES, DEFAULT_EXCLUDES
-from black.const import STDIN_PLACEHOLDER
-from black.nodes import STARS, syms, is_simple_decorator_expression
-from black.lines import Line, EmptyLineTracker
-from black.linegen import transform_line, LineGenerator, LN
+from _black_version import version as __version__
+from black.cache import Cache, filter_cached, get_cache_info, read_cache, write_cache
 from black.comments import normalize_fmt_off
-from black.mode import FUTURE_FLAG_TO_FEATURE, Mode, TargetVersion
-from black.mode import Feature, supports_feature, VERSION_TO_FEATURES
-from black.cache import read_cache, write_cache, get_cache_info, filter_cached, Cache
-from black.concurrency import cancel, shutdown, maybe_install_uvloop
-from black.output import dump_to_file, ipynb_diff, diff, color_diff, out, err
-from black.report import Report, Changed, NothingChanged
-from black.files import find_project_root, find_pyproject_toml, parse_pyproject_toml
-from black.files import gen_python_files, get_gitignore, normalize_path_maybe_ignore
-from black.files import wrap_stream_for_windows
-from black.parsing import InvalidInput  # noqa F401
-from black.parsing import lib2to3_parse, parse_ast, stringify_ast
+from black.concurrency import cancel, maybe_install_uvloop, shutdown
+from black.const import (
+    DEFAULT_EXCLUDES,
+    DEFAULT_INCLUDES,
+    DEFAULT_LINE_LENGTH,
+    STDIN_PLACEHOLDER,
+)
+from black.files import (
+    find_project_root,
+    find_pyproject_toml,
+    find_user_pyproject_toml,
+    gen_python_files,
+    get_gitignore,
+    normalize_path_maybe_ignore,
+    parse_pyproject_toml,
+    wrap_stream_for_windows,
+)
 from black.handle_ipynb_magics import (
-    mask_cell,
-    unmask_cell,
-    remove_trailing_semicolon,
-    put_trailing_semicolon_back,
-    TRANSFORMED_MAGICS,
     PYTHON_CELL_MAGICS,
+    TRANSFORMED_MAGICS,
     jupyter_dependencies_are_installed,
+    mask_cell,
+    put_trailing_semicolon_back,
+    remove_trailing_semicolon,
+    unmask_cell,
 )
-
-
-# lib2to3 fork
-from blib2to3.pytree import Node, Leaf
+from black.linegen import LN, LineGenerator, transform_line
+from black.lines import EmptyLineTracker, Line
+from black.mode import (
+    FUTURE_FLAG_TO_FEATURE,
+    VERSION_TO_FEATURES,
+    Feature,
+    Mode,
+    TargetVersion,
+    supports_feature,
+)
+from black.nodes import (
+    STARS,
+    is_number_token,
+    is_simple_decorator_expression,
+    is_string_token,
+    syms,
+)
+from black.output import color_diff, diff, dump_to_file, err, ipynb_diff, out
+from black.parsing import InvalidInput  # noqa F401
+from black.parsing import lib2to3_parse, parse_ast, stringify_ast
+from black.report import Changed, NothingChanged, Report
 from blib2to3.pgen2 import token
+from blib2to3.pytree import Leaf, Node
 
-from _black_version import version as __version__
+if TYPE_CHECKING:
+    from concurrent.futures import Executor
 
 COMPILED = Path(__file__).suffix in (".pyd", ".so")
 
@@ -223,6 +248,16 @@ def validate_regex(
         "(useful when piping source on standard input)."
     ),
 )
+@click.option(
+    "--python-cell-magics",
+    multiple=True,
+    help=(
+        "When processing Jupyter Notebooks, add the given magic to the list"
+        f" of known python-magics ({', '.join(PYTHON_CELL_MAGICS)})."
+        " Useful for formatting cells with custom python magics."
+    ),
+    default=[],
+)
 @click.option(
     "-S",
     "--skip-string-normalization",
@@ -239,9 +274,14 @@ def validate_regex(
     "--experimental-string-processing",
     is_flag=True,
     hidden=True,
+    help="(DEPRECATED and now included in --preview) Normalize string literals.",
+)
+@click.option(
+    "--preview",
+    is_flag=True,
     help=(
-        "Experimental option that performs more normalization on string literals."
-        " Currently disabled because it leads to some crashes."
+        "Enable potentially disruptive style changes that may be added to Black's main"
+        " functionality in the next major release."
     ),
 )
 @click.option(
@@ -273,7 +313,8 @@ def validate_regex(
     type=str,
     help=(
         "Require a specific version of Black to be running (useful for unifying results"
-        " across many environments e.g. with a pyproject.toml file)."
+        " across many environments e.g. with a pyproject.toml file). It can be"
+        " either a major version number or an exact version."
     ),
 )
 @click.option(
@@ -357,7 +398,10 @@ def validate_regex(
 )
 @click.version_option(
     version=__version__,
-    message=f"%(prog)s, %(version)s (compiled: {'yes' if COMPILED else 'no'})",
+    message=(
+        f"%(prog)s, %(version)s (compiled: {'yes' if COMPILED else 'no'})\n"
+        f"Python ({platform.python_implementation()}) {platform.python_version()}"
+    ),
 )
 @click.argument(
     "src",
@@ -383,7 +427,7 @@ def validate_regex(
     help="Read configuration from FILE path.",
 )
 @click.pass_context
-def main(
+def main(  # noqa: C901
     ctx: click.Context,
     code: Optional[str],
     line_length: int,
@@ -394,9 +438,11 @@ def main(
     fast: bool,
     pyi: bool,
     ipynb: bool,
+    python_cell_magics: Sequence[str],
     skip_string_normalization: bool,
     skip_magic_trailing_comma: bool,
     experimental_string_processing: bool,
+    preview: bool,
     quiet: bool,
     verbose: bool,
     required_version: Optional[str],
@@ -410,11 +456,65 @@ def main(
     config: Optional[str],
 ) -> None:
     """The uncompromising code formatter."""
-    if config and verbose:
-        out(f"Using configuration from {config}.", bold=False, fg="blue")
+    ctx.ensure_object(dict)
+
+    if src and code is not None:
+        out(
+            main.get_usage(ctx)
+            + "\n\n'SRC' and 'code' cannot be passed simultaneously."
+        )
+        ctx.exit(1)
+    if not src and code is None:
+        out(main.get_usage(ctx) + "\n\nOne of 'SRC' or 'code' is required.")
+        ctx.exit(1)
+
+    root, method = find_project_root(src) if code is None else (None, None)
+    ctx.obj["root"] = root
+
+    if verbose:
+        if root:
+            out(
+                f"Identified `{root}` as project root containing a {method}.",
+                fg="blue",
+            )
+
+            normalized = [
+                (normalize_path_maybe_ignore(Path(source), root), source)
+                for source in src
+            ]
+            srcs_string = ", ".join(
+                [
+                    f'"{_norm}"'
+                    if _norm
+                    else f'\033[31m"{source} (skipping - invalid)"\033[34m'
+                    for _norm, source in normalized
+                ]
+            )
+            out(f"Sources to be formatted: {srcs_string}", fg="blue")
+
+        if config:
+            config_source = ctx.get_parameter_source("config")
+            user_level_config = str(find_user_pyproject_toml())
+            if config == user_level_config:
+                out(
+                    "Using configuration from user-level config at "
+                    f"'{user_level_config}'.",
+                    fg="blue",
+                )
+            elif config_source in (
+                ParameterSource.DEFAULT,
+                ParameterSource.DEFAULT_MAP,
+            ):
+                out("Using configuration from project root.", fg="blue")
+            else:
+                out(f"Using configuration in '{config}'.", fg="blue")
 
     error_msg = "Oh no! 💥 💔 💥"
-    if required_version and required_version != __version__:
+    if (
+        required_version
+        and required_version != __version__
+        and required_version != __version__.split(".")[0]
+    ):
         err(
             f"{error_msg} The required version `{required_version}` does not match"
             f" the running version `{__version__}`!"
@@ -438,6 +538,8 @@ def main(
         string_normalization=not skip_string_normalization,
         magic_trailing_comma=not skip_magic_trailing_comma,
         experimental_string_processing=experimental_string_processing,
+        preview=preview,
+        python_cell_magics=set(python_cell_magics),
     )
 
     if code is not None:
@@ -495,6 +597,8 @@ def main(
             )
 
     if verbose or not quiet:
+        if code is None and (verbose or report.change_count or report.failure_count):
+            out()
         out(error_msg if report.return_code else "All done! ✨ 🍰 ✨")
         if code is None:
             click.echo(str(report), err=True)
@@ -515,14 +619,11 @@ def get_sources(
     stdin_filename: Optional[str],
 ) -> Set[Path]:
     """Compute the set of files to be formatted."""
-
-    root = find_project_root(src)
     sources: Set[Path] = set()
-    path_empty(src, "No Path provided. Nothing to do 😴", quiet, verbose, ctx)
 
     if exclude is None:
         exclude = re_compile_maybe_verbose(DEFAULT_EXCLUDES)
-        gitignore = get_gitignore(root)
+        gitignore = get_gitignore(ctx.obj["root"])
     else:
         gitignore = None
 
@@ -535,7 +636,7 @@ def get_sources(
             is_stdin = False
 
         if is_stdin or p.is_file():
-            normalized_path = normalize_path_maybe_ignore(p, root, report)
+            normalized_path = normalize_path_maybe_ignore(p, ctx.obj["root"], report)
             if normalized_path is None:
                 continue
 
@@ -562,7 +663,7 @@ def get_sources(
             sources.update(
                 gen_python_files(
                     p.iterdir(),
-                    root,
+                    ctx.obj["root"],
                     include,
                     exclude,
                     extend_exclude,
@@ -616,6 +717,9 @@ def reformat_code(
         report.failed(path, str(exc))
 
 
+# diff-shades depends on being to monkeypatch this function to operate. I know it's
+# not ideal, but this shouldn't cause any issues ... hopefully. ~ichard26
+@mypyc_attr(patchable=True)
 def reformat_one(
     src: Path, fast: bool, write_back: WriteBack, mode: Mode, report: "Report"
 ) -> None:
@@ -679,8 +783,9 @@ def reformat_many(
     workers: Optional[int],
 ) -> None:
     """Reformat multiple files using a ProcessPoolExecutor."""
+    from concurrent.futures import Executor, ProcessPoolExecutor, ThreadPoolExecutor
+
     executor: Executor
-    loop = asyncio.get_event_loop()
     worker_count = workers if workers is not None else DEFAULT_WORKERS
     if sys.platform == "win32":
         # Work around https://bugs.python.org/issue26903
@@ -695,6 +800,8 @@ def reformat_many(
         # any good due to the Global Interpreter Lock)
         executor = ThreadPoolExecutor(max_workers=1)
 
+    loop = asyncio.new_event_loop()
+    asyncio.set_event_loop(loop)
     try:
         loop.run_until_complete(
             schedule_formatting(
@@ -708,7 +815,10 @@ def reformat_many(
             )
         )
     finally:
-        shutdown(loop)
+        try:
+            shutdown(loop)
+        finally:
+            asyncio.set_event_loop(None)
         if executor is not None:
             executor.shutdown()
 
@@ -720,7 +830,7 @@ async def schedule_formatting(
     mode: Mode,
     report: "Report",
     loop: asyncio.AbstractEventLoop,
-    executor: Executor,
+    executor: "Executor",
 ) -> None:
     """Run formatting of `sources` in parallel using the provided `executor`.
 
@@ -907,17 +1017,7 @@ def check_stability_and_equivalence(
     content differently.
     """
     assert_equivalent(src_contents, dst_contents)
-
-    # Forced second pass to work around optional trailing commas (becoming
-    # forced trailing commas on pass 2) interacting differently with optional
-    # parentheses.  Admittedly ugly.
-    dst_contents_pass2 = format_str(dst_contents, mode=mode)
-    if dst_contents != dst_contents_pass2:
-        dst_contents = dst_contents_pass2
-        assert_equivalent(src_contents, dst_contents, pass_num=2)
-        assert_stable(src_contents, dst_contents, mode=mode)
-    # Note: no need to explicitly call `assert_stable` if `dst_contents` was
-    # the same as `dst_contents_pass2`.
+    assert_stable(src_contents, dst_contents, mode=mode)
 
 
 def format_file_contents(src_contents: str, *, fast: bool, mode: Mode) -> FileContent:
@@ -943,7 +1043,7 @@ def format_file_contents(src_contents: str, *, fast: bool, mode: Mode) -> FileCo
     return dst_contents
 
 
-def validate_cell(src: str) -> None:
+def validate_cell(src: str, mode: Mode) -> None:
     """Check that cell does not already contain TransformerManager transformations,
     or non-Python cell magics, which might cause tokenizer_rt to break because of
     indentations.
@@ -962,7 +1062,10 @@ def validate_cell(src: str) -> None:
     """
     if any(transformed_magic in src for transformed_magic in TRANSFORMED_MAGICS):
         raise NothingChanged
-    if src[:2] == "%%" and src.split()[0][2:] not in PYTHON_CELL_MAGICS:
+    if (
+        src[:2] == "%%"
+        and src.split()[0][2:] not in PYTHON_CELL_MAGICS | mode.python_cell_magics
+    ):
         raise NothingChanged
 
 
@@ -982,7 +1085,7 @@ def format_cell(src: str, *, fast: bool, mode: Mode) -> str:
     could potentially be automagics or multi-line magics, which
     are currently not supported.
     """
-    validate_cell(src)
+    validate_cell(src, mode)
     src_without_trailing_semicolon, has_trailing_semicolon = remove_trailing_semicolon(
         src
     )
@@ -1044,7 +1147,7 @@ def format_ipynb_string(src_contents: str, *, fast: bool, mode: Mode) -> FileCon
         raise NothingChanged
 
 
-def format_str(src_contents: str, *, mode: Mode) -> FileContent:
+def format_str(src_contents: str, *, mode: Mode) -> str:
     """Reformat a string and return new contents.
 
     `mode` determines formatting options, such as how many characters per line are
@@ -1074,28 +1177,26 @@ def format_str(src_contents: str, *, mode: Mode) -> FileContent:
         hey
 
     """
+    dst_contents = _format_str_once(src_contents, mode=mode)
+    # Forced second pass to work around optional trailing commas (becoming
+    # forced trailing commas on pass 2) interacting differently with optional
+    # parentheses.  Admittedly ugly.
+    if src_contents != dst_contents:
+        return _format_str_once(dst_contents, mode=mode)
+    return dst_contents
+
+
+def _format_str_once(src_contents: str, *, mode: Mode) -> str:
     src_node = lib2to3_parse(src_contents.lstrip(), mode.target_versions)
     dst_contents = []
-    future_imports = get_future_imports(src_node)
     if mode.target_versions:
         versions = mode.target_versions
     else:
+        future_imports = get_future_imports(src_node)
         versions = detect_target_versions(src_node, future_imports=future_imports)
 
-    # TODO: fully drop support and this code hopefully in January 2022 :D
-    if TargetVersion.PY27 in mode.target_versions or versions == {TargetVersion.PY27}:
-        msg = (
-            "DEPRECATION: Python 2 support will be removed in the first stable release "
-            "expected in January 2022."
-        )
-        err(msg, fg="yellow", bold=True)
-
-    normalize_fmt_off(src_node)
-    lines = LineGenerator(
-        mode=mode,
-        remove_u_prefix="unicode_literals" in future_imports
-        or supports_feature(versions, Feature.UNICODE_LITERALS),
-    )
+    normalize_fmt_off(src_node, preview=mode.preview)
+    lines = LineGenerator(mode=mode)
     elt = EmptyLineTracker(is_pyi=mode.is_pyi)
     empty_line = Line(mode=mode)
     after = 0
@@ -1156,23 +1257,14 @@ def get_features_used(  # noqa: C901
         }
 
     for n in node.pre_order():
-        if n.type == token.STRING:
-            value_head = n.value[:2]  # type: ignore
+        if is_string_token(n):
+            value_head = n.value[:2]
             if value_head in {'f"', 'F"', "f'", "F'", "rf", "fr", "RF", "FR"}:
                 features.add(Feature.F_STRINGS)
 
-        elif n.type == token.NUMBER:
-            assert isinstance(n, Leaf)
+        elif is_number_token(n):
             if "_" in n.value:
                 features.add(Feature.NUMERIC_UNDERSCORES)
-            elif n.value.endswith(("L", "l")):
-                # Python 2: 10L
-                features.add(Feature.LONG_INT_LITERAL)
-            elif len(n.value) >= 2 and n.value[0] == "0" and n.value[1].isdigit():
-                # Python 2: 0123; 00123; ...
-                if not all(char == "0" for char in n.value):
-                    # although we don't want to match 0000 or similar
-                    features.add(Feature.OCTAL_INT_LITERAL)
 
         elif n.type == token.SLASH:
             if n.parent and n.parent.type in {
@@ -1210,31 +1302,39 @@ def get_features_used(  # noqa: C901
                         if argch.type in STARS:
                             features.add(feature)
 
-        # Python 2 only features (for its deprecation) except for integers, see above
-        elif n.type == syms.print_stmt:
-            features.add(Feature.PRINT_STMT)
-        elif n.type == syms.exec_stmt:
-            features.add(Feature.EXEC_STMT)
-        elif n.type == syms.tfpdef:
-            # def set_position((x, y), value):
-            #     ...
-            features.add(Feature.AUTOMATIC_PARAMETER_UNPACKING)
-        elif n.type == syms.except_clause:
-            # try:
-            #     ...
-            # except Exception, err:
-            #     ...
-            if len(n.children) >= 4:
-                if n.children[-2].type == token.COMMA:
-                    features.add(Feature.COMMA_STYLE_EXCEPT)
-        elif n.type == syms.raise_stmt:
-            # raise Exception, "msg"
-            if len(n.children) >= 4:
-                if n.children[-2].type == token.COMMA:
-                    features.add(Feature.COMMA_STYLE_RAISE)
-        elif n.type == token.BACKQUOTE:
-            # `i'm surprised this ever existed`
-            features.add(Feature.BACKQUOTE_REPR)
+        elif (
+            n.type in {syms.return_stmt, syms.yield_expr}
+            and len(n.children) >= 2
+            and n.children[1].type == syms.testlist_star_expr
+            and any(child.type == syms.star_expr for child in n.children[1].children)
+        ):
+            features.add(Feature.UNPACKING_ON_FLOW)
+
+        elif (
+            n.type == syms.annassign
+            and len(n.children) >= 4
+            and n.children[3].type == syms.testlist_star_expr
+        ):
+            features.add(Feature.ANN_ASSIGN_EXTENDED_RHS)
+
+        elif (
+            n.type == syms.except_clause
+            and len(n.children) >= 2
+            and n.children[1].type == token.STAR
+        ):
+            features.add(Feature.EXCEPT_STAR)
+
+        elif n.type in {syms.subscriptlist, syms.trailer} and any(
+            child.type == syms.star_expr for child in n.children
+        ):
+            features.add(Feature.VARIADIC_GENERICS)
+
+        elif (
+            n.type == syms.tname_star
+            and len(n.children) == 3
+            and n.children[2].type == syms.star_expr
+        ):
+            features.add(Feature.VARIADIC_GENERICS)
 
     return features
 
@@ -1299,13 +1399,16 @@ def get_future_imports(node: Node) -> Set[str]:
     return imports
 
 
-def assert_equivalent(src: str, dst: str, *, pass_num: int = 1) -> None:
+def assert_equivalent(src: str, dst: str) -> None:
     """Raise AssertionError if `src` and `dst` aren't equivalent."""
     try:
         src_ast = parse_ast(src)
     except Exception as exc:
         raise AssertionError(
-            f"cannot use --safe with this file; failed to parse source file: {exc}"
+            "cannot use --safe with this file; failed to parse source file AST: "
+            f"{exc}\n"
+            "This could be caused by running Black with an older Python version "
+            "that does not support new syntax used in your source file."
         ) from exc
 
     try:
@@ -1313,7 +1416,7 @@ def assert_equivalent(src: str, dst: str, *, pass_num: int = 1) -> None:
     except Exception as exc:
         log = dump_to_file("".join(traceback.format_tb(exc.__traceback__)), dst)
         raise AssertionError(
-            f"INTERNAL ERROR: Black produced invalid code on pass {pass_num}: {exc}. "
+            f"INTERNAL ERROR: Black produced invalid code: {exc}. "
             "Please report a bug on https://github.com/psf/black/issues.  "
             f"This invalid output might be helpful: {log}"
         ) from None
@@ -1324,14 +1427,17 @@ def assert_equivalent(src: str, dst: str, *, pass_num: int = 1) -> None:
         log = dump_to_file(diff(src_ast_str, dst_ast_str, "src", "dst"))
         raise AssertionError(
             "INTERNAL ERROR: Black produced code that is not equivalent to the"
-            f" source on pass {pass_num}.  Please report a bug on "
+            " source.  Please report a bug on "
             f"https://github.com/psf/black/issues.  This diff might be helpful: {log}"
         ) from None
 
 
 def assert_stable(src: str, dst: str, mode: Mode) -> None:
     """Raise AssertionError if `dst` reformats differently the second time."""
-    newdst = format_str(dst, mode=mode)
+    # We shouldn't call format_str() here, because that formats the string
+    # twice and may hide a bug where we bounce back and forth between two
+    # versions.
+    newdst = _format_str_once(dst, mode=mode)
     if dst != newdst:
         log = dump_to_file(
             str(mode),
@@ -1365,13 +1471,23 @@ def patch_click() -> None:
     file paths is minimal since it's Python source code.  Moreover, this crash was
     spurious on Python 3.7 thanks to PEP 538 and PEP 540.
     """
+    modules: List[Any] = []
     try:
         from click import core
-        from click import _unicodefun
-    except ModuleNotFoundError:
-        return
+    except ImportError:
+        pass
+    else:
+        modules.append(core)
+    try:
+        # Removed in Click 8.1.0 and newer; we keep this around for users who have
+        # older versions installed.
+        from click import _unicodefun  # type: ignore
+    except ImportError:
+        pass
+    else:
+        modules.append(_unicodefun)
 
-    for module in (core, _unicodefun):
+    for module in modules:
         if hasattr(module, "_verify_python3_env"):
             module._verify_python3_env = lambda: None  # type: ignore
         if hasattr(module, "_verify_python_env"):