]> 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:

Use better default venv dir when using neovim (#937)
[etc/vim.git] / black.py
index 05edf1a2ae320c3bd3b9a11c57b2edd3a314a4cf..65545e9f21d69dea541eec99f5bfa85cf4cf44e6 100644 (file)
--- a/black.py
+++ b/black.py
@@ -12,7 +12,7 @@ from multiprocessing import Manager, freeze_support
 import os
 from pathlib import Path
 import pickle
-import re
+import regex as re
 import signal
 import sys
 import tempfile
@@ -51,12 +51,10 @@ from blib2to3.pgen2 import driver, token
 from blib2to3.pgen2.grammar import Grammar
 from blib2to3.pgen2.parse import ParseError
 
+from _version import version as __version__
 
-__version__ = "19.3b0"
 DEFAULT_LINE_LENGTH = 88
-DEFAULT_EXCLUDES = (
-    r"/(\.eggs|\.git|\.hg|\.mypy_cache|\.nox|\.tox|\.venv|_build|buck-out|build|dist)/"
-)
+DEFAULT_EXCLUDES = r"/(\.eggs|\.git|\.hg|\.mypy_cache|\.nox|\.tox|\.venv|\.svn|_build|buck-out|build|dist)/"  # noqa: B950
 DEFAULT_INCLUDES = r"\.pyi?$"
 CACHE_DIR = Path(user_cache_dir("black", version=__version__))
 
@@ -433,6 +431,7 @@ def main(
     report = Report(check=check, quiet=quiet, verbose=verbose)
     root = find_project_root(src)
     sources: Set[Path] = set()
+    path_empty(src, quiet, verbose, ctx)
     for s in src:
         p = Path(s)
         if p.is_dir():
@@ -446,7 +445,7 @@ def main(
             err(f"invalid path: {s}")
     if len(sources) == 0:
         if verbose or not quiet:
-            out("No paths given. Nothing to do 😴")
+            out("No Python files are present to be formatted. Nothing to do 😴")
         ctx.exit(0)
 
     if len(sources) == 1:
@@ -468,6 +467,16 @@ def main(
     ctx.exit(report.return_code)
 
 
+def path_empty(src: Tuple[str], quiet: bool, verbose: bool, ctx: click.Context) -> None:
+    """
+    Exit if there is no `src` provided for formatting
+    """
+    if not src:
+        if verbose or not quiet:
+            out("No Path provided. Nothing to do 😴")
+            ctx.exit(0)
+
+
 def reformat_one(
     src: Path, fast: bool, write_back: WriteBack, mode: FileMode, report: "Report"
 ) -> None:
@@ -1229,6 +1238,61 @@ class Line:
             Leaf(token.DOT, ".") for _ in range(3)
         ]
 
+    @property
+    def is_collection_with_optional_trailing_comma(self) -> bool:
+        """Is this line a collection literal with a trailing comma that's optional?
+
+        Note that the trailing comma in a 1-tuple is not optional.
+        """
+        if not self.leaves or len(self.leaves) < 4:
+            return False
+        # Look for and address a trailing colon.
+        if self.leaves[-1].type == token.COLON:
+            closer = self.leaves[-2]
+            close_index = -2
+        else:
+            closer = self.leaves[-1]
+            close_index = -1
+        if closer.type not in CLOSING_BRACKETS or self.inside_brackets:
+            return False
+        if closer.type == token.RPAR:
+            # Tuples require an extra check, because if there's only
+            # one element in the tuple removing the comma unmakes the
+            # tuple.
+            #
+            # We also check for parens before looking for the trailing
+            # comma because in some cases (eg assigning a dict
+            # literal) the literal gets wrapped in temporary parens
+            # during parsing. This case is covered by the
+            # collections.py test data.
+            opener = closer.opening_bracket
+            for _open_index, leaf in enumerate(self.leaves):
+                if leaf is opener:
+                    break
+            else:
+                # Couldn't find the matching opening paren, play it safe.
+                return False
+            commas = 0
+            comma_depth = self.leaves[close_index - 1].bracket_depth
+            for leaf in self.leaves[_open_index + 1 : close_index]:
+                if leaf.bracket_depth == comma_depth and leaf.type == token.COMMA:
+                    commas += 1
+            if commas > 1:
+                # We haven't looked yet for the trailing comma because
+                # we might also have caught noop parens.
+                return self.leaves[close_index - 1].type == token.COMMA
+            elif commas == 1:
+                return False  # it's either a one-tuple or didn't have a trailing comma
+            if self.leaves[close_index - 1].type in CLOSING_BRACKETS:
+                close_index -= 1
+                closer = self.leaves[close_index]
+                if closer.type == token.RPAR:
+                    # TODO: this is a gut feeling. Will we ever see this?
+                    return False
+        if self.leaves[close_index - 1].type != token.COMMA:
+            return False
+        return True
+
     @property
     def is_def(self) -> bool:
         """Is this a function definition? (Also returns True for async defs.)"""
@@ -1281,7 +1345,7 @@ class Line:
                     return True
         return False
 
-    def contains_inner_type_comments(self) -> bool:
+    def contains_uncollapsable_type_comments(self) -> bool:
         ignored_ids = set()
         try:
             last_leaf = self.leaves[-1]
@@ -1298,13 +1362,50 @@ class Line:
         except IndexError:
             return False
 
+        # A type comment is uncollapsable if it is attached to a leaf
+        # that isn't at the end of the line (since that could cause it
+        # to get associated to a different argument) or if there are
+        # comments before it (since that could cause it to get hidden
+        # behind a comment.
+        comment_seen = False
         for leaf_id, comments in self.comments.items():
-            if leaf_id in ignored_ids:
-                continue
-
             for comment in comments:
                 if is_type_comment(comment):
-                    return True
+                    if leaf_id not in ignored_ids or comment_seen:
+                        return True
+
+                comment_seen = True
+
+        return False
+
+    def contains_unsplittable_type_ignore(self) -> bool:
+        if not self.leaves:
+            return False
+
+        # If a 'type: ignore' is attached to the end of a line, we
+        # can't split the line, because we can't know which of the
+        # subexpressions the ignore was meant to apply to.
+        #
+        # We only want this to apply to actual physical lines from the
+        # original source, though: we don't want the presence of a
+        # 'type: ignore' at the end of a multiline expression to
+        # justify pushing it all onto one line. Thus we
+        # (unfortunately) need to check the actual source lines and
+        # only report an unsplittable 'type: ignore' if this line was
+        # one line in the original code.
+
+        # Grab the first and last line numbers, skipping generated leaves
+        first_line = next((l.lineno for l in self.leaves if l.lineno != 0), 0)
+        last_line = next((l.lineno for l in reversed(self.leaves) if l.lineno != 0), 0)
+
+        if first_line == last_line:
+            # We look at the last two leaves since a comma or an
+            # invisible paren could have been added at the end of the
+            # line.
+            for node in self.leaves[-2:]:
+                for comment in self.comments.get(id(node), []):
+                    if is_type_comment(comment, " ignore"):
+                        return True
 
         return False
 
@@ -1317,60 +1418,39 @@ class Line:
 
     def maybe_remove_trailing_comma(self, closing: Leaf) -> bool:
         """Remove trailing comma if there is one and it's safe."""
+        if not (self.leaves and self.leaves[-1].type == token.COMMA):
+            return False
+        # We remove trailing commas only in the case of importing a
+        # single name from a module.
         if not (
             self.leaves
+            and self.is_import
+            and len(self.leaves) > 4
             and self.leaves[-1].type == token.COMMA
             and closing.type in CLOSING_BRACKETS
+            and self.leaves[-4].type == token.NAME
+            and (
+                # regular `from foo import bar,`
+                self.leaves[-4].value == "import"
+                # `from foo import (bar as baz,)
+                or (
+                    len(self.leaves) > 6
+                    and self.leaves[-6].value == "import"
+                    and self.leaves[-3].value == "as"
+                )
+                # `from foo import bar as baz,`
+                or (
+                    len(self.leaves) > 5
+                    and self.leaves[-5].value == "import"
+                    and self.leaves[-3].value == "as"
+                )
+            )
+            and closing.type == token.RPAR
         ):
             return False
 
-        if closing.type == token.RBRACE:
-            self.remove_trailing_comma()
-            return True
-
-        if closing.type == token.RSQB:
-            comma = self.leaves[-1]
-            if comma.parent and comma.parent.type == syms.listmaker:
-                self.remove_trailing_comma()
-                return True
-
-        # For parens let's check if it's safe to remove the comma.
-        # Imports are always safe.
-        if self.is_import:
-            self.remove_trailing_comma()
-            return True
-
-        # Otherwise, if the trailing one is the only one, we might mistakenly
-        # change a tuple into a different type by removing the comma.
-        depth = closing.bracket_depth + 1
-        commas = 0
-        opening = closing.opening_bracket
-        for _opening_index, leaf in enumerate(self.leaves):
-            if leaf is opening:
-                break
-
-        else:
-            return False
-
-        for leaf in self.leaves[_opening_index + 1 :]:
-            if leaf is closing:
-                break
-
-            bracket_depth = leaf.bracket_depth
-            if bracket_depth == depth and leaf.type == token.COMMA:
-                commas += 1
-                if leaf.parent and leaf.parent.type in {
-                    syms.arglist,
-                    syms.typedargslist,
-                }:
-                    commas += 1
-                    break
-
-        if commas > 1:
-            self.remove_trailing_comma()
-            return True
-
-        return False
+        self.remove_trailing_comma()
+        return True
 
     def append_comment(self, comment: Leaf) -> bool:
         """Add an inline or standalone comment to the line."""
@@ -2313,9 +2393,13 @@ def split_line(
     line_str = str(line).strip("\n")
 
     if (
-        not line.contains_inner_type_comments()
+        not line.contains_uncollapsable_type_comments()
         and not line.should_explode
-        and is_line_short_enough(line, line_length=line_length, line_str=line_str)
+        and not line.is_collection_with_optional_trailing_comma
+        and (
+            is_line_short_enough(line, line_length=line_length, line_str=line_str)
+            or line.contains_unsplittable_type_ignore()
+        )
     ):
         yield line
         return
@@ -2688,12 +2772,14 @@ def is_import(leaf: Leaf) -> bool:
     )
 
 
-def is_type_comment(leaf: Leaf) -> bool:
+def is_type_comment(leaf: Leaf, suffix: str = "") -> bool:
     """Return True if the given leaf is a special comment.
     Only returns true for type comments for now."""
     t = leaf.type
     v = leaf.value
-    return t in {token.COMMENT, t == STANDALONE_COMMENT} and v.startswith("# type:")
+    return t in {token.COMMENT, t == STANDALONE_COMMENT} and v.startswith(
+        "# type:" + suffix
+    )
 
 
 def normalize_prefix(leaf: Leaf, *, inside_brackets: bool) -> None:
@@ -2878,7 +2964,15 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
             if is_walrus_assignment(child):
                 continue
             if child.type == syms.atom:
-                if maybe_make_parens_invisible_in_atom(child, parent=node):
+                # Determines if the underlying atom should be surrounded with
+                # invisible params - also makes parens invisible recursively
+                # within the atom and removes repeated invisible parens within
+                # the atom
+                should_surround_with_parens = maybe_make_parens_invisible_in_atom(
+                    child, parent=node
+                )
+
+                if should_surround_with_parens:
                     lpar = Leaf(token.LPAR, "")
                     rpar = Leaf(token.RPAR, "")
                     index = child.remove() or 0
@@ -2995,6 +3089,8 @@ def generate_ignored_nodes(leaf: Leaf) -> Iterator[LN]:
 
 def maybe_make_parens_invisible_in_atom(node: LN, parent: LN) -> bool:
     """If it's safe, make the parens in the atom `node` invisible, recursively.
+    Additionally, remove repeated, adjacent invisible parens from the atom `node`
+    as they are redundant.
 
     Returns whether the node should itself be wrapped in invisible parentheses.
 
@@ -3011,15 +3107,40 @@ def maybe_make_parens_invisible_in_atom(node: LN, parent: LN) -> bool:
     first = node.children[0]
     last = node.children[-1]
     if first.type == token.LPAR and last.type == token.RPAR:
+        middle = node.children[1]
         # make parentheses invisible
         first.value = ""  # type: ignore
         last.value = ""  # type: ignore
-        maybe_make_parens_invisible_in_atom(node.children[1], parent=parent)
+        maybe_make_parens_invisible_in_atom(middle, parent=parent)
+
+        if is_atom_with_invisible_parens(middle):
+            # Strip the invisible parens from `middle` by replacing
+            # it with the child in-between the invisible parens
+            middle.replace(middle.children[1])
+
         return False
 
     return True
 
 
+def is_atom_with_invisible_parens(node: LN) -> bool:
+    """Given a `LN`, determines whether it's an atom `node` with invisible
+    parens. Useful in dedupe-ing and normalizing parens.
+    """
+    if isinstance(node, Leaf) or node.type != syms.atom:
+        return False
+
+    first, last = node.children[0], node.children[-1]
+    return (
+        isinstance(first, Leaf)
+        and first.type == token.LPAR
+        and first.value == ""
+        and isinstance(last, Leaf)
+        and last.type == token.RPAR
+        and last.value == ""
+    )
+
+
 def is_empty_tuple(node: LN) -> bool:
     """Return True if `node` holds an empty tuple."""
     return (
@@ -3718,7 +3839,8 @@ def re_compile_maybe_verbose(regex: str) -> Pattern[str]:
     """
     if "\n" in regex:
         regex = "(?x)" + regex
-    return re.compile(regex)
+    compiled: Pattern[str] = re.compile(regex)
+    return compiled
 
 
 def enumerate_reversed(sequence: Sequence[T]) -> Iterator[Tuple[Index, T]]: