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

Bump sphinx from 4.3.2 to 4.4.0 in /docs (#2776)
[etc/vim.git] / src / black / parsing.py
index 8e9feea9120db796afdf980b10ddc5e7864cebd0..13fa67ee84dabb07f183f8bee0bd8b8742fac490 100644 (file)
@@ -2,23 +2,35 @@
 Parse Python code and perform AST validation.
 """
 import ast
+import platform
 import sys
-from typing import Iterable, Iterator, List, Set, Union
+from typing import Any, Iterable, Iterator, List, Set, Tuple, Type, Union
+
+if sys.version_info < (3, 8):
+    from typing_extensions import Final
+else:
+    from typing import Final
 
 # lib2to3 fork
 from blib2to3.pytree import Node, Leaf
-from blib2to3 import pygram, pytree
+from blib2to3 import pygram
 from blib2to3.pgen2 import driver
 from blib2to3.pgen2.grammar import Grammar
 from blib2to3.pgen2.parse import ParseError
+from blib2to3.pgen2.tokenize import TokenError
 
 from black.mode import TargetVersion, Feature, supports_feature
 from black.nodes import syms
 
+ast3: Any
+
+_IS_PYPY = platform.python_implementation() == "PyPy"
+
 try:
-    from typed_ast import ast3, ast27
+    from typed_ast import ast3
 except ImportError:
-    if sys.version_info < (3, 8):
+    # Either our python version is too low, or we're on pypy
+    if sys.version_info < (3, 7) or (sys.version_info < (3, 8) and not _IS_PYPY):
         print(
             "The typed_ast package is required but not installed.\n"
             "You can upgrade to Python 3.8+ or install typed_ast with\n"
@@ -27,7 +39,11 @@ except ImportError:
         )
         sys.exit(1)
     else:
-        ast3 = ast27 = ast
+        ast3 = ast
+
+
+PY310_HINT: Final = "Consider using --target-version py310 to parse Python 3.10 code."
+PY2_HINT: Final = "Python 2 support was removed in version 22.0."
 
 
 class InvalidInput(ValueError):
@@ -42,26 +58,17 @@ def get_grammars(target_versions: Set[TargetVersion]) -> List[Grammar]:
             pygram.python_grammar_no_print_statement_no_exec_statement_async_keywords,
             # Python 3.0-3.6
             pygram.python_grammar_no_print_statement_no_exec_statement,
-            # Python 2.7 with future print_function import
-            pygram.python_grammar_no_print_statement,
-            # Python 2.7
-            pygram.python_grammar,
-        ]
-
-    if all(version.is_python2() for version in target_versions):
-        # Python 2-only code, so try Python 2 grammars.
-        return [
-            # Python 2.7 with future print_function import
-            pygram.python_grammar_no_print_statement,
-            # Python 2.7
-            pygram.python_grammar,
         ]
 
-    # Python 3-compatible code, so only try Python 3 grammar.
     grammars = []
+    if supports_feature(target_versions, Feature.PATTERN_MATCHING):
+        # Python 3.10+
+        grammars.append(pygram.python_grammar_soft_keywords)
     # If we have to parse both, try to parse async as a keyword first
-    if not supports_feature(target_versions, Feature.ASYNC_IDENTIFIERS):
-        # Python 3.7+
+    if not supports_feature(
+        target_versions, Feature.ASYNC_IDENTIFIERS
+    ) and not supports_feature(target_versions, Feature.PATTERN_MATCHING):
+        # Python 3.7-3.9
         grammars.append(
             pygram.python_grammar_no_print_statement_no_exec_statement_async_keywords
         )
@@ -78,8 +85,9 @@ def lib2to3_parse(src_txt: str, target_versions: Iterable[TargetVersion] = ()) -
     if not src_txt.endswith("\n"):
         src_txt += "\n"
 
-    for grammar in get_grammars(set(target_versions)):
-        drv = driver.Driver(grammar, pytree.convert)
+    grammars = get_grammars(set(target_versions))
+    for grammar in grammars:
+        drv = driver.Driver(grammar)
         try:
             result = drv.parse_string(src_txt, True)
             break
@@ -92,7 +100,27 @@ def lib2to3_parse(src_txt: str, target_versions: Iterable[TargetVersion] = ()) -
             except IndexError:
                 faulty_line = "<line number missing in source>"
             exc = InvalidInput(f"Cannot parse: {lineno}:{column}: {faulty_line}")
+
+        except TokenError as te:
+            # In edge cases these are raised; and typically don't have a "faulty_line".
+            lineno, column = te.args[1]
+            exc = InvalidInput(f"Cannot parse: {lineno}:{column}: {te.args[0]}")
+
     else:
+        if pygram.python_grammar_soft_keywords not in grammars and matches_grammar(
+            src_txt, pygram.python_grammar_soft_keywords
+        ):
+            original_msg = exc.args[0]
+            msg = f"{original_msg}\n{PY310_HINT}"
+            raise InvalidInput(msg) from None
+
+        if matches_grammar(src_txt, pygram.python_grammar) or matches_grammar(
+            src_txt, pygram.python_grammar_no_print_statement
+        ):
+            original_msg = exc.args[0]
+            msg = f"{original_msg}\n{PY2_HINT}"
+            raise InvalidInput(msg) from None
+
         raise exc from None
 
     if isinstance(result, Leaf):
@@ -100,52 +128,82 @@ def lib2to3_parse(src_txt: str, target_versions: Iterable[TargetVersion] = ()) -
     return result
 
 
+def matches_grammar(src_txt: str, grammar: Grammar) -> bool:
+    drv = driver.Driver(grammar)
+    try:
+        drv.parse_string(src_txt, True)
+    except (ParseError, TokenError, IndentationError):
+        return False
+    else:
+        return True
+
+
 def lib2to3_unparse(node: Node) -> str:
     """Given a lib2to3 node, return its string representation."""
     code = str(node)
     return code
 
 
-def parse_ast(src: str) -> Union[ast.AST, ast3.AST, ast27.AST]:
+def parse_single_version(
+    src: str, version: Tuple[int, int]
+) -> Union[ast.AST, ast3.AST]:
     filename = "<unknown>"
-    if sys.version_info >= (3, 8):
-        # TODO: support Python 4+ ;)
-        for minor_version in range(sys.version_info[1], 4, -1):
-            try:
-                return ast.parse(src, filename, feature_version=(3, minor_version))
-            except SyntaxError:
-                continue
-    else:
-        for feature_version in (7, 6):
-            try:
-                return ast3.parse(src, filename, feature_version=feature_version)
-            except SyntaxError:
-                continue
-    if ast27.__name__ == "ast":
-        raise SyntaxError(
-            "The requested source code has invalid Python 3 syntax.\n"
-            "If you are trying to format Python 2 files please reinstall Black"
-            " with the 'python2' extra: `python3 -m pip install black[python2]`."
-        )
-    return ast27.parse(src)
+    # typed_ast is needed because of feature version limitations in the builtin ast
+    if sys.version_info >= (3, 8) and version >= (3,):
+        return ast.parse(src, filename, feature_version=version)
+    elif version >= (3,):
+        if _IS_PYPY:
+            return ast3.parse(src, filename)
+        else:
+            return ast3.parse(src, filename, feature_version=version[1])
+    raise AssertionError("INTERNAL ERROR: Tried parsing unsupported Python version!")
+
+
+def parse_ast(src: str) -> Union[ast.AST, ast3.AST]:
+    # TODO: support Python 4+ ;)
+    versions = [(3, minor) for minor in range(3, sys.version_info[1] + 1)]
+
+    first_error = ""
+    for version in sorted(versions, reverse=True):
+        try:
+            return parse_single_version(src, version)
+        except SyntaxError as e:
+            if not first_error:
+                first_error = str(e)
+
+    raise SyntaxError(first_error)
+
+
+ast3_AST: Final[Type[ast3.AST]] = ast3.AST
 
 
-def stringify_ast(
-    node: Union[ast.AST, ast3.AST, ast27.AST], depth: int = 0
-) -> Iterator[str]:
+def _normalize(lineend: str, value: str) -> str:
+    # To normalize, we strip any leading and trailing space from
+    # each line...
+    stripped: List[str] = [i.strip() for i in value.splitlines()]
+    normalized = lineend.join(stripped)
+    # ...and remove any blank lines at the beginning and end of
+    # the whole string
+    return normalized.strip()
+
+
+def stringify_ast(node: Union[ast.AST, ast3.AST], depth: int = 0) -> Iterator[str]:
     """Simple visitor generating strings to compare ASTs by content."""
 
     node = fixup_ast_constants(node)
 
     yield f"{'  ' * depth}{node.__class__.__name__}("
 
+    type_ignore_classes: Tuple[Type[Any], ...]
     for field in sorted(node._fields):  # noqa: F402
-        # TypeIgnore has only one field 'lineno' which breaks this comparison
-        type_ignore_classes = (ast3.TypeIgnore, ast27.TypeIgnore)
-        if sys.version_info >= (3, 8):
-            type_ignore_classes += (ast.TypeIgnore,)
-        if isinstance(node, type_ignore_classes):
-            break
+        # TypeIgnore will not be present using pypy < 3.8, so need for this
+        if not (_IS_PYPY and sys.version_info < (3, 8)):
+            # TypeIgnore has only one field 'lineno' which breaks this comparison
+            type_ignore_classes = (ast3.TypeIgnore,)
+            if sys.version_info >= (3, 8):
+                type_ignore_classes += (ast.TypeIgnore,)
+            if isinstance(node, type_ignore_classes):
+                break
 
         try:
             value = getattr(node, field)
@@ -160,38 +218,34 @@ def stringify_ast(
                 # parentheses and they change the AST.
                 if (
                     field == "targets"
-                    and isinstance(node, (ast.Delete, ast3.Delete, ast27.Delete))
-                    and isinstance(item, (ast.Tuple, ast3.Tuple, ast27.Tuple))
+                    and isinstance(node, (ast.Delete, ast3.Delete))
+                    and isinstance(item, (ast.Tuple, ast3.Tuple))
                 ):
                     for item in item.elts:
                         yield from stringify_ast(item, depth + 2)
 
-                elif isinstance(item, (ast.AST, ast3.AST, ast27.AST)):
+                elif isinstance(item, (ast.AST, ast3.AST)):
                     yield from stringify_ast(item, depth + 2)
 
-        elif isinstance(value, (ast.AST, ast3.AST, ast27.AST)):
+        # Note that we are referencing the typed-ast ASTs via global variables and not
+        # direct module attribute accesses because that breaks mypyc. It's probably
+        # something to do with the ast3 variables being marked as Any leading
+        # mypy to think this branch is always taken, leaving the rest of the code
+        # unanalyzed. Tighting up the types for the typed-ast AST types avoids the
+        # mypyc crash.
+        elif isinstance(value, (ast.AST, ast3_AST)):
             yield from stringify_ast(value, depth + 2)
 
         else:
             # Constant strings may be indented across newlines, if they are
             # docstrings; fold spaces after newlines when comparing. Similarly,
             # trailing and leading space may be removed.
-            # Note that when formatting Python 2 code, at least with Windows
-            # line-endings, docstrings can end up here as bytes instead of
-            # str so make sure that we handle both cases.
             if (
                 isinstance(node, ast.Constant)
                 and field == "value"
-                and isinstance(value, (str, bytes))
+                and isinstance(value, str)
             ):
-                lineend = "\n" if isinstance(value, str) else b"\n"
-                # To normalize, we strip any leading and trailing space from
-                # each line...
-                stripped = [line.strip() for line in value.splitlines()]
-                normalized = lineend.join(stripped)  # type: ignore[attr-defined]
-                # ...and remove any blank lines at the beginning and end of
-                # the whole string
-                normalized = normalized.strip()
+                normalized = _normalize("\n", value)
             else:
                 normalized = value
             yield f"{'  ' * (depth+2)}{normalized!r},  # {value.__class__.__name__}"
@@ -199,14 +253,12 @@ def stringify_ast(
     yield f"{'  ' * depth})  # /{node.__class__.__name__}"
 
 
-def fixup_ast_constants(
-    node: Union[ast.AST, ast3.AST, ast27.AST]
-) -> Union[ast.AST, ast3.AST, ast27.AST]:
+def fixup_ast_constants(node: Union[ast.AST, ast3.AST]) -> Union[ast.AST, ast3.AST]:
     """Map ast nodes deprecated in 3.8 to Constant."""
-    if isinstance(node, (ast.Str, ast3.Str, ast27.Str, ast.Bytes, ast3.Bytes)):
+    if isinstance(node, (ast.Str, ast3.Str, ast.Bytes, ast3.Bytes)):
         return ast.Constant(value=node.s)
 
-    if isinstance(node, (ast.Num, ast3.Num, ast27.Num)):
+    if isinstance(node, (ast.Num, ast3.Num)):
         return ast.Constant(value=node.n)
 
     if isinstance(node, (ast.NameConstant, ast3.NameConstant)):