]> 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:

Remove deprecated --py36 option (#1236)
[etc/vim.git] / black.py
index 40b4d2b12876013b0925c9a1c706110befeb9b28..3ab4bc76bb23836a1f12bbe4953a5fb6bcbd7b55 100644 (file)
--- a/black.py
+++ b/black.py
@@ -1,6 +1,8 @@
 import ast
 import asyncio
-from concurrent.futures import Executor, ProcessPoolExecutor
+from abc import ABC, abstractmethod
+from collections import defaultdict
+from concurrent.futures import Executor, ThreadPoolExecutor, ProcessPoolExecutor
 from contextlib import contextmanager
 from datetime import datetime
 from enum import Enum
@@ -32,17 +34,23 @@ from typing import (
     Pattern,
     Sequence,
     Set,
+    Sized,
     Tuple,
+    Type,
     TypeVar,
     Union,
     cast,
+    TYPE_CHECKING,
 )
+from typing_extensions import Final
+from mypy_extensions import mypyc_attr
 
 from appdirs import user_cache_dir
-from attr import dataclass, evolve, Factory
+from dataclasses import dataclass, field, replace
 import click
 import toml
 from typed_ast import ast3, ast27
+from pathspec import PathSpec
 
 # lib2to3 fork
 from blib2to3.pytree import Node, Leaf, type_repr
@@ -51,15 +59,18 @@ from blib2to3.pgen2 import driver, token
 from blib2to3.pgen2.grammar import Grammar
 from blib2to3.pgen2.parse import ParseError
 
-from _version import version as __version__
+from _black_version import version as __version__
+
+if TYPE_CHECKING:
+    import colorama  # noqa: F401
 
 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__))
 
+STRING_PREFIX_CHARS: Final = "furbFURB"  # All possible string prefix characters.
+
 
 # types
 FileContent = str
@@ -67,11 +78,13 @@ Encoding = str
 NewLine = str
 Depth = int
 NodeType = int
+ParserState = int
 LeafID = int
+StringID = int
 Priority = int
 Index = int
 LN = Union[Leaf, Node]
-SplitFunc = Callable[["Line", Collection["Feature"]], Iterator["Line"]]
+Transformer = Callable[["Line", Collection["Feature"]], Iterator["Line"]]
 Timestamp = float
 FileSize = int
 CacheInfo = Tuple[Timestamp, FileSize]
@@ -87,7 +100,11 @@ class NothingChanged(UserWarning):
     """Raised when reformatted code is the same as source."""
 
 
-class CannotSplit(Exception):
+class CannotTransform(Exception):
+    """Base class for errors raised by Transformers."""
+
+
+class CannotSplit(CannotTransform):
     """A readable split that fits the allotted line length is impossible."""
 
 
@@ -95,17 +112,51 @@ class InvalidInput(ValueError):
     """Raised when input source code fails all parse attempts."""
 
 
+T = TypeVar("T")
+E = TypeVar("E", bound=Exception)
+
+
+class Ok(Generic[T]):
+    def __init__(self, value: T) -> None:
+        self._value = value
+
+    def ok(self) -> T:
+        return self._value
+
+
+class Err(Generic[E]):
+    def __init__(self, e: E) -> None:
+        self._e = e
+
+    def err(self) -> E:
+        return self._e
+
+
+# The 'Result' return type is used to implement an error-handling model heavily
+# influenced by that used by the Rust programming language
+# (see https://doc.rust-lang.org/book/ch09-00-error-handling.html).
+Result = Union[Ok[T], Err[E]]
+TResult = Result[T, CannotTransform]  # (T)ransform Result
+TMatchResult = TResult[Index]
+
+
 class WriteBack(Enum):
     NO = 0
     YES = 1
     DIFF = 2
     CHECK = 3
+    COLOR_DIFF = 4
 
     @classmethod
-    def from_configuration(cls, *, check: bool, diff: bool) -> "WriteBack":
+    def from_configuration(
+        cls, *, check: bool, diff: bool, color: bool = False
+    ) -> "WriteBack":
         if check and not diff:
             return cls.CHECK
 
+        if diff and color:
+            return cls.COLOR_DIFF
+
         return cls.DIFF if diff else cls.YES
 
 
@@ -185,8 +236,8 @@ VERSION_TO_FEATURES: Dict[TargetVersion, Set[Feature]] = {
 
 
 @dataclass
-class FileMode:
-    target_versions: Set[TargetVersion] = Factory(set)
+class Mode:
+    target_versions: Set[TargetVersion] = field(default_factory=set)
     line_length: int = DEFAULT_LINE_LENGTH
     string_normalization: bool = True
     is_pyi: bool = False
@@ -208,30 +259,46 @@ class FileMode:
         return ".".join(parts)
 
 
+# Legacy name, left for integrations.
+FileMode = Mode
+
+
 def supports_feature(target_versions: Set[TargetVersion], feature: Feature) -> bool:
     return all(feature in VERSION_TO_FEATURES[version] for version in target_versions)
 
 
+def find_pyproject_toml(path_search_start: str) -> Optional[str]:
+    """Find the absolute filepath to a pyproject.toml if it exists"""
+    path_project_root = find_project_root(path_search_start)
+    path_pyproject_toml = path_project_root / "pyproject.toml"
+    return str(path_pyproject_toml) if path_pyproject_toml.is_file() else None
+
+
+def parse_pyproject_toml(path_config: str) -> Dict[str, Any]:
+    """Parse a pyproject toml file, pulling out relevant parts for Black
+
+    If parsing fails, will raise a toml.TomlDecodeError
+    """
+    pyproject_toml = toml.load(path_config)
+    config = pyproject_toml.get("tool", {}).get("black", {})
+    return {k.replace("--", "").replace("-", "_"): v for k, v in config.items()}
+
+
 def read_pyproject_toml(
-    ctx: click.Context, param: click.Parameter, value: Union[str, int, bool, None]
+    ctx: click.Context, param: click.Parameter, value: Optional[str]
 ) -> Optional[str]:
     """Inject Black configuration from "pyproject.toml" into defaults in `ctx`.
 
     Returns the path to a successfully found and read configuration file, None
     otherwise.
     """
-    assert not isinstance(value, (int, bool)), "Invalid parameter type passed"
     if not value:
-        root = find_project_root(ctx.params.get("src", ()))
-        path = root / "pyproject.toml"
-        if path.is_file():
-            value = str(path)
-        else:
+        value = find_pyproject_toml(ctx.params.get("src", ()))
+        if value is None:
             return None
 
     try:
-        pyproject_toml = toml.load(value)
-        config = pyproject_toml.get("tool", {}).get("black", {})
+        config = parse_pyproject_toml(value)
     except (toml.TomlDecodeError, OSError) as e:
         raise click.FileError(
             filename=value, hint=f"Error reading configuration file: {e}"
@@ -240,14 +307,32 @@ def read_pyproject_toml(
     if not config:
         return None
 
-    if ctx.default_map is None:
-        ctx.default_map = {}
-    ctx.default_map.update(  # type: ignore  # bad types in .pyi
-        {k.replace("--", "").replace("-", "_"): v for k, v in config.items()}
-    )
+    target_version = config.get("target_version")
+    if target_version is not None and not isinstance(target_version, list):
+        raise click.BadOptionUsage(
+            "target-version", f"Config key target-version must be a list"
+        )
+
+    default_map: Dict[str, Any] = {}
+    if ctx.default_map:
+        default_map.update(ctx.default_map)
+    default_map.update(config)
+
+    ctx.default_map = default_map
     return value
 
 
+def target_version_option_callback(
+    c: click.Context, p: Union[click.Option, click.Parameter], v: Tuple[str, ...]
+) -> List[TargetVersion]:
+    """Compute the target versions from a --target-version flag.
+
+    This is its own function because mypy couldn't infer the type correctly
+    when it was a lambda, causing mypyc trouble.
+    """
+    return [TargetVersion[val.upper()] for val in v]
+
+
 @click.command(context_settings=dict(help_option_names=["-h", "--help"]))
 @click.option("-c", "--code", type=str, help="Format the code passed in as a string.")
 @click.option(
@@ -262,29 +347,19 @@ def read_pyproject_toml(
     "-t",
     "--target-version",
     type=click.Choice([v.name.lower() for v in TargetVersion]),
-    callback=lambda c, p, v: [TargetVersion[val.upper()] for val in v],
+    callback=target_version_option_callback,
     multiple=True,
     help=(
-        "Python versions that should be supported by Black's output. [default: "
-        "per-file auto-detection]"
-    ),
-)
-@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. Deprecated; use --target-version instead. "
-        "[default: per-file auto-detection]"
+        "Python versions that should be supported by Black's output. [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)."
+        "Format all input files like typing stubs regardless of file extension (useful"
+        " when piping source on standard input)."
     ),
 )
 @click.option(
@@ -297,9 +372,9 @@ def read_pyproject_toml(
     "--check",
     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."
+        "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(
@@ -307,6 +382,11 @@ def read_pyproject_toml(
     is_flag=True,
     help="Don't write the files back, just output a diff for each file on stdout.",
 )
+@click.option(
+    "--color/--no-color",
+    is_flag=True,
+    help="Show colored diff. Only applies when `--diff` is given.",
+)
 @click.option(
     "--fast/--safe",
     is_flag=True,
@@ -317,11 +397,10 @@ def read_pyproject_toml(
     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."
+        "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,
 )
@@ -330,20 +409,28 @@ def read_pyproject_toml(
     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."
+        "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(
+    "--force-exclude",
+    type=str,
+    help=(
+        "Like --exclude, but files and directories matching this regex will be "
+        "excluded even when they are passed explicitly as arguments"
+    ),
+)
 @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."
+        "Don't emit non-error messages to stderr. Errors are still emitted; silence"
+        " those with 2>/dev/null."
     ),
 )
 @click.option(
@@ -351,8 +438,8 @@ def read_pyproject_toml(
     "--verbose",
     is_flag=True,
     help=(
-        "Also emit messages to stderr about files that were not changed or were "
-        "ignored due to --exclude=."
+        "Also emit messages to stderr about files that were not changed or were ignored"
+        " due to --exclude=."
     ),
 )
 @click.version_option(version=__version__)
@@ -367,7 +454,12 @@ def read_pyproject_toml(
 @click.option(
     "--config",
     type=click.Path(
-        exists=False, file_okay=True, dir_okay=False, readable=True, allow_dash=False
+        exists=True,
+        file_okay=True,
+        dir_okay=False,
+        readable=True,
+        allow_dash=False,
+        path_type=str,
     ),
     is_eager=True,
     callback=read_pyproject_toml,
@@ -381,35 +473,26 @@ def main(
     target_version: List[TargetVersion],
     check: bool,
     diff: bool,
+    color: bool,
     fast: bool,
     pyi: bool,
-    py36: bool,
     skip_string_normalization: bool,
     quiet: bool,
     verbose: bool,
     include: str,
     exclude: str,
-    src: Tuple[str],
+    force_exclude: Optional[str],
+    src: Tuple[str, ...],
     config: Optional[str],
 ) -> None:
     """The uncompromising code formatter."""
-    write_back = WriteBack.from_configuration(check=check, diff=diff)
+    write_back = WriteBack.from_configuration(check=check, diff=diff, color=color)
     if target_version:
-        if py36:
-            err(f"Cannot use both --target-version and --py36")
-            ctx.exit(2)
-        else:
-            versions = set(target_version)
-    elif py36:
-        err(
-            "--py36 is deprecated and will be removed in a future version. "
-            "Use --target-version py36 instead."
-        )
-        versions = PY36_VERSIONS
+        versions = set(target_version)
     else:
         # We'll autodetect later.
         versions = set()
-    mode = FileMode(
+    mode = Mode(
         target_versions=versions,
         line_length=line_length,
         is_pyi=pyi,
@@ -420,6 +503,57 @@ def main(
     if code is not None:
         print(format_str(code, mode=mode))
         ctx.exit(0)
+    report = Report(check=check, diff=diff, quiet=quiet, verbose=verbose)
+    sources = get_sources(
+        ctx=ctx,
+        src=src,
+        quiet=quiet,
+        verbose=verbose,
+        include=include,
+        exclude=exclude,
+        force_exclude=force_exclude,
+        report=report,
+    )
+
+    path_empty(
+        sources,
+        "No Python files are present to be formatted. Nothing to do 😴",
+        quiet,
+        verbose,
+        ctx,
+    )
+
+    if len(sources) == 1:
+        reformat_one(
+            src=sources.pop(),
+            fast=fast,
+            write_back=write_back,
+            mode=mode,
+            report=report,
+        )
+    else:
+        reformat_many(
+            sources=sources, fast=fast, write_back=write_back, mode=mode, report=report
+        )
+
+    if verbose or not quiet:
+        out("Oh no! 💥 💔 💥" if report.return_code else "All done! ✨ 🍰 ✨")
+        click.secho(str(report), err=True)
+    ctx.exit(report.return_code)
+
+
+def get_sources(
+    *,
+    ctx: click.Context,
+    src: Tuple[str, ...],
+    quiet: bool,
+    verbose: bool,
+    include: str,
+    exclude: str,
+    force_exclude: Optional[str],
+    report: "Report",
+) -> Set[Path]:
+    """Compute the set of files to be formatted."""
     try:
         include_regex = re_compile_maybe_verbose(include)
     except re.error:
@@ -430,57 +564,61 @@ def main(
     except re.error:
         err(f"Invalid regular expression for exclude given: {exclude!r}")
         ctx.exit(2)
-    report = Report(check=check, quiet=quiet, verbose=verbose)
+    try:
+        force_exclude_regex = (
+            re_compile_maybe_verbose(force_exclude) if force_exclude else None
+        )
+    except re.error:
+        err(f"Invalid regular expression for force_exclude given: {force_exclude!r}")
+        ctx.exit(2)
+
     root = find_project_root(src)
     sources: Set[Path] = set()
-    path_empty(src, quiet, verbose, ctx)
+    path_empty(src, "No Path provided. Nothing to do 😴", quiet, verbose, ctx)
+    exclude_regexes = [exclude_regex]
+    if force_exclude_regex is not None:
+        exclude_regexes.append(force_exclude_regex)
+
     for s in src:
         p = Path(s)
         if p.is_dir():
             sources.update(
-                gen_python_files_in_dir(p, root, include_regex, exclude_regex, report)
+                gen_python_files(
+                    p.iterdir(),
+                    root,
+                    include_regex,
+                    exclude_regexes,
+                    report,
+                    get_gitignore(root),
+                )
             )
-        elif p.is_file() or s == "-":
-            # if a file was explicitly given, we don't care about its extension
+        elif s == "-":
             sources.add(p)
+        elif p.is_file():
+            sources.update(
+                gen_python_files(
+                    [p], root, None, exclude_regexes, report, get_gitignore(root)
+                )
+            )
         else:
             err(f"invalid path: {s}")
-    if len(sources) == 0:
-        if verbose or not quiet:
-            out("No Python files are present to be formatted. Nothing to do 😴")
-        ctx.exit(0)
-
-    if len(sources) == 1:
-        reformat_one(
-            src=sources.pop(),
-            fast=fast,
-            write_back=write_back,
-            mode=mode,
-            report=report,
-        )
-    else:
-        reformat_many(
-            sources=sources, fast=fast, write_back=write_back, mode=mode, report=report
-        )
-
-    if verbose or not quiet:
-        out("Oh no! 💥 💔 💥" if report.return_code else "All done! ✨ 🍰 ✨")
-        click.secho(str(report), err=True)
-    ctx.exit(report.return_code)
+    return sources
 
 
-def path_empty(src: Tuple[str], quiet: bool, verbose: bool, ctx: click.Context) -> None:
+def path_empty(
+    src: Sized, msg: str, quiet: bool, verbose: bool, ctx: click.Context
+) -> None:
     """
     Exit if there is no `src` provided for formatting
     """
-    if not src:
+    if len(src) == 0:
         if verbose or not quiet:
-            out("No Path provided. Nothing to do 😴")
+            out(msg)
             ctx.exit(0)
 
 
 def reformat_one(
-    src: Path, fast: bool, write_back: WriteBack, mode: FileMode, report: "Report"
+    src: Path, fast: bool, write_back: WriteBack, mode: Mode, report: "Report"
 ) -> None:
     """Reformat a single file under `src` without spawning child processes.
 
@@ -513,19 +651,24 @@ def reformat_one(
 
 
 def reformat_many(
-    sources: Set[Path],
-    fast: bool,
-    write_back: WriteBack,
-    mode: FileMode,
-    report: "Report",
+    sources: Set[Path], fast: bool, write_back: WriteBack, mode: Mode, report: "Report"
 ) -> None:
     """Reformat multiple files using a ProcessPoolExecutor."""
+    executor: Executor
     loop = asyncio.get_event_loop()
     worker_count = os.cpu_count()
     if sys.platform == "win32":
         # Work around https://bugs.python.org/issue26903
         worker_count = min(worker_count, 61)
-    executor = ProcessPoolExecutor(max_workers=worker_count)
+    try:
+        executor = ProcessPoolExecutor(max_workers=worker_count)
+    except OSError:
+        # we arrive here if the underlying system does not support multi-processing
+        # like in AWS Lambda, in which case we gracefully fallback to
+        # a ThreadPollExecutor with just a single worker (more workers would not do us
+        # any good due to the Global Interpreter Lock)
+        executor = ThreadPoolExecutor(max_workers=1)
+
     try:
         loop.run_until_complete(
             schedule_formatting(
@@ -540,14 +683,15 @@ def reformat_many(
         )
     finally:
         shutdown(loop)
-        executor.shutdown()
+        if executor is not None:
+            executor.shutdown()
 
 
 async def schedule_formatting(
     sources: Set[Path],
     fast: bool,
     write_back: WriteBack,
-    mode: FileMode,
+    mode: Mode,
     report: "Report",
     loop: asyncio.AbstractEventLoop,
     executor: Executor,
@@ -584,7 +728,7 @@ async def schedule_formatting(
         ): src
         for src in sorted(sources)
     }
-    pending: Iterable[asyncio.Future] = tasks.keys()
+    pending: Iterable["asyncio.Future[bool]"] = tasks.keys()
     try:
         loop.add_signal_handler(signal.SIGINT, cancel, pending)
         loop.add_signal_handler(signal.SIGTERM, cancel, pending)
@@ -617,7 +761,7 @@ async def schedule_formatting(
 def format_file_in_place(
     src: Path,
     fast: bool,
-    mode: FileMode,
+    mode: Mode,
     write_back: WriteBack = WriteBack.NO,
     lock: Any = None,  # multiprocessing.Manager().Lock() is some crazy proxy
 ) -> bool:
@@ -628,7 +772,7 @@ def format_file_in_place(
     `mode` and `fast` options are passed to :func:`format_file_contents`.
     """
     if src.suffix == ".pyi":
-        mode = evolve(mode, is_pyi=True)
+        mode = replace(mode, is_pyi=True)
 
     then = datetime.utcfromtimestamp(src.stat().st_mtime)
     with open(src, "rb") as buf:
@@ -638,15 +782,18 @@ def format_file_in_place(
     except NothingChanged:
         return False
 
-    if write_back == write_back.YES:
+    if write_back == WriteBack.YES:
         with open(src, "w", encoding=encoding, newline=newline) as f:
             f.write(dst_contents)
-    elif write_back == write_back.DIFF:
+    elif write_back in (WriteBack.DIFF, WriteBack.COLOR_DIFF):
         now = datetime.utcnow()
         src_name = f"{src}\t{then} +0000"
         dst_name = f"{src}\t{now} +0000"
         diff_contents = diff(src_contents, dst_contents, src_name, dst_name)
 
+        if write_back == write_back.COLOR_DIFF:
+            diff_contents = color_diff(diff_contents)
+
         with lock or nullcontext():
             f = io.TextIOWrapper(
                 sys.stdout.buffer,
@@ -654,14 +801,59 @@ def format_file_in_place(
                 newline=newline,
                 write_through=True,
             )
+            f = wrap_stream_for_windows(f)
             f.write(diff_contents)
             f.detach()
 
     return True
 
 
+def color_diff(contents: str) -> str:
+    """Inject the ANSI color codes to the diff."""
+    lines = contents.split("\n")
+    for i, line in enumerate(lines):
+        if line.startswith("+++") or line.startswith("---"):
+            line = "\033[1;37m" + line + "\033[0m"  # bold white, reset
+        if line.startswith("@@"):
+            line = "\033[36m" + line + "\033[0m"  # cyan, reset
+        if line.startswith("+"):
+            line = "\033[32m" + line + "\033[0m"  # green, reset
+        elif line.startswith("-"):
+            line = "\033[31m" + line + "\033[0m"  # red, reset
+        lines[i] = line
+    return "\n".join(lines)
+
+
+def wrap_stream_for_windows(
+    f: io.TextIOWrapper,
+) -> Union[io.TextIOWrapper, "colorama.AnsiToWin32.AnsiToWin32"]:
+    """
+    Wrap the stream in colorama's wrap_stream so colors are shown on Windows.
+
+    If `colorama` is not found, then no change is made. If `colorama` does
+    exist, then it handles the logic to determine whether or not to change
+    things.
+    """
+    try:
+        from colorama import initialise
+
+        # We set `strip=False` so that we can don't have to modify
+        # test_express_diff_with_color.
+        f = initialise.wrap_stream(
+            f, convert=None, strip=False, autoreset=False, wrap=True
+        )
+
+        # wrap_stream returns a `colorama.AnsiToWin32.AnsiToWin32` object
+        # which does not have a `detach()` method. So we fake one.
+        f.detach = lambda *args, **kwargs: None  # type: ignore
+    except ImportError:
+        pass
+
+    return f
+
+
 def format_stdin_to_stdout(
-    fast: bool, *, write_back: WriteBack = WriteBack.NO, mode: FileMode
+    fast: bool, *, write_back: WriteBack = WriteBack.NO, mode: Mode
 ) -> bool:
     """Format file on stdin. Return True if changed.
 
@@ -685,17 +877,19 @@ def format_stdin_to_stdout(
         )
         if write_back == WriteBack.YES:
             f.write(dst)
-        elif write_back == WriteBack.DIFF:
+        elif write_back in (WriteBack.DIFF, WriteBack.COLOR_DIFF):
             now = datetime.utcnow()
             src_name = f"STDIN\t{then} +0000"
             dst_name = f"STDOUT\t{now} +0000"
-            f.write(diff(src, dst, src_name, dst_name))
+            d = diff(src, dst, src_name, dst_name)
+            if write_back == WriteBack.COLOR_DIFF:
+                d = color_diff(d)
+                f = wrap_stream_for_windows(f)
+            f.write(d)
         f.detach()
 
 
-def format_file_contents(
-    src_contents: str, *, fast: bool, mode: FileMode
-) -> FileContent:
+def format_file_contents(src_contents: str, *, fast: bool, mode: Mode) -> FileContent:
     """Reformat contents a file and return new contents.
 
     If `fast` is False, additionally confirm that the reformatted code is
@@ -715,11 +909,34 @@ def format_file_contents(
     return dst_contents
 
 
-def format_str(src_contents: str, *, mode: FileMode) -> FileContent:
+def format_str(src_contents: str, *, mode: Mode) -> FileContent:
     """Reformat a string and return new contents.
 
     `mode` determines formatting options, such as how many characters per line are
-    allowed.
+    allowed.  Example:
+
+    >>> import black
+    >>> print(black.format_str("def f(arg:str='')->None:...", mode=Mode()))
+    def f(arg: str = "") -> None:
+        ...
+
+    A more complex example:
+    >>> print(
+    ...   black.format_str(
+    ...     "def f(arg:str='')->None: hey",
+    ...     mode=black.Mode(
+    ...       target_versions={black.TargetVersion.PY36},
+    ...       line_length=10,
+    ...       string_normalization=False,
+    ...       is_pyi=False,
+    ...     ),
+    ...   ),
+    ... )
+    def f(
+        arg: str = '',
+    ) -> None:
+        hey
+
     """
     src_node = lib2to3_parse(src_contents.lstrip(), mode.target_versions)
     dst_contents = []
@@ -744,13 +961,14 @@ def format_str(src_contents: str, *, mode: FileMode) -> FileContent:
         if supports_feature(versions, feature)
     }
     for current_line in lines.visit(src_node):
-        for _ in range(after):
-            dst_contents.append(str(empty_line))
+        dst_contents.append(str(empty_line) * after)
         before, after = elt.maybe_empty_lines(current_line)
-        for _ in range(before):
-            dst_contents.append(str(empty_line))
-        for line in split_line(
-            current_line, line_length=mode.line_length, features=split_line_features
+        dst_contents.append(str(empty_line) * before)
+        for line in transform_line(
+            current_line,
+            line_length=mode.line_length,
+            normalize_strings=mode.string_normalization,
+            features=split_line_features,
         ):
             dst_contents.append(str(line))
     return "".join(dst_contents)
@@ -786,7 +1004,8 @@ def get_grammars(target_versions: Set[TargetVersion]) -> List[Grammar]:
             # Python 2.7
             pygram.python_grammar,
         ]
-    elif all(version.is_python2() for version in target_versions):
+
+    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
@@ -794,21 +1013,21 @@ def get_grammars(target_versions: Set[TargetVersion]) -> List[Grammar]:
             # Python 2.7
             pygram.python_grammar,
         ]
-    else:
-        # Python 3-compatible code, so only try Python 3 grammar.
-        grammars = []
-        # 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+
-            grammars.append(
-                pygram.python_grammar_no_print_statement_no_exec_statement_async_keywords  # noqa: B950
-            )
-        if not supports_feature(target_versions, Feature.ASYNC_KEYWORDS):
-            # Python 3.0-3.6
-            grammars.append(pygram.python_grammar_no_print_statement_no_exec_statement)
-        # At least one of the above branches must have been taken, because every Python
-        # version has exactly one of the two 'ASYNC_*' flags
-        return grammars
+
+    # Python 3-compatible code, so only try Python 3 grammar.
+    grammars = []
+    # 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+
+        grammars.append(
+            pygram.python_grammar_no_print_statement_no_exec_statement_async_keywords
+        )
+    if not supports_feature(target_versions, Feature.ASYNC_KEYWORDS):
+        # Python 3.0-3.6
+        grammars.append(pygram.python_grammar_no_print_statement_no_exec_statement)
+    # At least one of the above branches must have been taken, because every Python
+    # version has exactly one of the two 'ASYNC_*' flags
+    return grammars
 
 
 def lib2to3_parse(src_txt: str, target_versions: Iterable[TargetVersion] = ()) -> Node:
@@ -844,9 +1063,6 @@ def lib2to3_unparse(node: Node) -> str:
     return code
 
 
-T = TypeVar("T")
-
-
 class Visitor(Generic[T]):
     """Basic lib2to3 visitor that yields things of type `T` on `visit()`."""
 
@@ -863,8 +1079,16 @@ class Visitor(Generic[T]):
         if node.type < 256:
             name = token.tok_name[node.type]
         else:
-            name = type_repr(node.type)
-        yield from getattr(self, f"visit_{name}", self.visit_default)(node)
+            name = str(type_repr(node.type))
+        # We explicitly branch on whether a visitor exists (instead of
+        # using self.visit_default as the default arg to getattr) in order
+        # to save needing to create a bound method object and so mypyc can
+        # generate a native call to visit_default.
+        visitf = getattr(self, f"visit_{name}", None)
+        if visitf:
+            yield from visitf(node)
+        else:
+            yield from self.visit_default(node)
 
     def visit_default(self, node: LN) -> Iterator[T]:
         """Default `visit_*()` implementation. Recurses to children of `node`."""
@@ -909,8 +1133,8 @@ class DebugVisitor(Visitor[T]):
         list(v.visit(code))
 
 
-WHITESPACE = {token.DEDENT, token.INDENT, token.NEWLINE}
-STATEMENT = {
+WHITESPACE: Final = {token.DEDENT, token.INDENT, token.NEWLINE}
+STATEMENT: Final = {
     syms.if_stmt,
     syms.while_stmt,
     syms.for_stmt,
@@ -920,10 +1144,10 @@ STATEMENT = {
     syms.funcdef,
     syms.classdef,
 }
-STANDALONE_COMMENT = 153
+STANDALONE_COMMENT: Final = 153
 token.tok_name[STANDALONE_COMMENT] = "STANDALONE_COMMENT"
-LOGIC_OPERATORS = {"and", "or"}
-COMPARATORS = {
+LOGIC_OPERATORS: Final = {"and", "or"}
+COMPARATORS: Final = {
     token.LESS,
     token.GREATER,
     token.EQEQUAL,
@@ -931,7 +1155,7 @@ COMPARATORS = {
     token.LESSEQUAL,
     token.GREATEREQUAL,
 }
-MATH_OPERATORS = {
+MATH_OPERATORS: Final = {
     token.VBAR,
     token.CIRCUMFLEX,
     token.AMPER,
@@ -947,23 +1171,23 @@ MATH_OPERATORS = {
     token.TILDE,
     token.DOUBLESTAR,
 }
-STARS = {token.STAR, token.DOUBLESTAR}
-VARARGS_SPECIALS = STARS | {token.SLASH}
-VARARGS_PARENTS = {
+STARS: Final = {token.STAR, token.DOUBLESTAR}
+VARARGS_SPECIALS: Final = STARS | {token.SLASH}
+VARARGS_PARENTS: Final = {
     syms.arglist,
     syms.argument,  # double star in arglist
     syms.trailer,  # single argument to call
     syms.typedargslist,
     syms.varargslist,  # lambdas
 }
-UNPACKING_PARENTS = {
+UNPACKING_PARENTS: Final = {
     syms.atom,  # single element of a list or set literal
     syms.dictsetmaker,
     syms.listmaker,
     syms.testlist_gexp,
     syms.testlist_star_expr,
 }
-TEST_DESCENDANTS = {
+TEST_DESCENDANTS: Final = {
     syms.test,
     syms.lambdef,
     syms.or_test,
@@ -980,7 +1204,7 @@ TEST_DESCENDANTS = {
     syms.term,
     syms.power,
 }
-ASSIGNMENTS = {
+ASSIGNMENTS: Final = {
     "=",
     "+=",
     "-=",
@@ -996,13 +1220,13 @@ ASSIGNMENTS = {
     "**=",
     "//=",
 }
-COMPREHENSION_PRIORITY = 20
-COMMA_PRIORITY = 18
-TERNARY_PRIORITY = 16
-LOGIC_PRIORITY = 14
-STRING_PRIORITY = 12
-COMPARATOR_PRIORITY = 10
-MATH_PRIORITIES = {
+COMPREHENSION_PRIORITY: Final = 20
+COMMA_PRIORITY: Final = 18
+TERNARY_PRIORITY: Final = 16
+LOGIC_PRIORITY: Final = 14
+STRING_PRIORITY: Final = 12
+COMPARATOR_PRIORITY: Final = 10
+MATH_PRIORITIES: Final = {
     token.VBAR: 9,
     token.CIRCUMFLEX: 8,
     token.AMPER: 7,
@@ -1018,7 +1242,7 @@ MATH_PRIORITIES = {
     token.TILDE: 3,
     token.DOUBLESTAR: 2,
 }
-DOT_PRIORITY = 1
+DOT_PRIORITY: Final = 1
 
 
 @dataclass
@@ -1026,11 +1250,11 @@ class BracketTracker:
     """Keeps track of brackets on a line."""
 
     depth: int = 0
-    bracket_match: Dict[Tuple[Depth, NodeType], Leaf] = Factory(dict)
-    delimiters: Dict[LeafID, Priority] = Factory(dict)
+    bracket_match: Dict[Tuple[Depth, NodeType], Leaf] = field(default_factory=dict)
+    delimiters: Dict[LeafID, Priority] = field(default_factory=dict)
     previous: Optional[Leaf] = None
-    _for_loop_depths: List[int] = Factory(list)
-    _lambda_argument_depths: List[int] = Factory(list)
+    _for_loop_depths: List[int] = field(default_factory=list)
+    _lambda_argument_depths: List[int] = field(default_factory=list)
 
     def mark(self, leaf: Leaf) -> None:
         """Mark `leaf` with bracket-related metadata. Keep track of delimiters.
@@ -1158,9 +1382,10 @@ class Line:
     """Holds leaves and comments. Can be printed with `str(line)`."""
 
     depth: int = 0
-    leaves: List[Leaf] = Factory(list)
-    comments: Dict[LeafID, List[Leaf]] = Factory(dict)  # keys ordered like `leaves`
-    bracket_tracker: BracketTracker = Factory(BracketTracker)
+    leaves: List[Leaf] = field(default_factory=list)
+    # keys ordered like `leaves`
+    comments: Dict[LeafID, List[Leaf]] = field(default_factory=dict)
+    bracket_tracker: BracketTracker = field(default_factory=BracketTracker)
     inside_brackets: bool = False
     should_explode: bool = False
 
@@ -1248,6 +1473,7 @@ class Line:
         """
         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]
@@ -1257,6 +1483,7 @@ class Line:
             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
@@ -1271,9 +1498,11 @@ class Line:
             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]:
@@ -1283,16 +1512,20 @@ class Line:
                 # 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
@@ -1342,9 +1575,9 @@ class Line:
     def contains_standalone_comments(self, depth_limit: int = sys.maxsize) -> bool:
         """If so, needs to be split before emitting."""
         for leaf in self.leaves:
-            if leaf.type == STANDALONE_COMMENT:
-                if leaf.bracket_depth <= depth_limit:
-                    return True
+            if leaf.type == STANDALONE_COMMENT and leaf.bracket_depth <= depth_limit:
+                return True
+
         return False
 
     def contains_uncollapsable_type_comments(self) -> bool:
@@ -1373,7 +1606,10 @@ class Line:
         for leaf_id, comments in self.comments.items():
             for comment in comments:
                 if is_type_comment(comment):
-                    if leaf_id not in ignored_ids or comment_seen:
+                    if comment_seen or (
+                        not is_type_comment(comment, " ignore")
+                        and leaf_id not in ignored_ids
+                    ):
                         return True
 
                 comment_seen = True
@@ -1412,16 +1648,13 @@ class Line:
         return False
 
     def contains_multiline_strings(self) -> bool:
-        for leaf in self.leaves:
-            if is_multiline_string(leaf):
-                return True
-
-        return False
+        return any(is_multiline_string(leaf) for leaf in self.leaves)
 
     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 (
@@ -1486,6 +1719,7 @@ class Line:
                 comment.type = STANDALONE_COMMENT
                 comment.prefix = ""
                 return False
+
             last_leaf = self.leaves[-2]
         self.comments.setdefault(id(last_leaf), []).append(comment)
         return True
@@ -1520,6 +1754,13 @@ class Line:
             n.type in TEST_DESCENDANTS for n in subscript_start.pre_order()
         )
 
+    def clone(self) -> "Line":
+        return Line(
+            depth=self.depth,
+            inside_brackets=self.inside_brackets,
+            should_explode=self.should_explode,
+        )
+
     def __str__(self) -> str:
         """Render the line."""
         if not self:
@@ -1533,6 +1774,7 @@ class Line:
             res += str(leaf)
         for comment in itertools.chain.from_iterable(self.comments.values()):
             res += str(comment)
+
         return res + "\n"
 
     def __bool__(self) -> bool:
@@ -1553,7 +1795,7 @@ class EmptyLineTracker:
     is_pyi: bool = False
     previous_line: Optional[Line] = None
     previous_after: int = 0
-    previous_defs: List[int] = Factory(list)
+    previous_defs: List[int] = field(default_factory=list)
 
     def maybe_empty_lines(self, current_line: Line) -> Tuple[int, int]:
         """Return the number of extra empty lines before and after the `current_line`.
@@ -1667,7 +1909,7 @@ class LineGenerator(Visitor[Line]):
 
     is_pyi: bool = False
     normalize_strings: bool = True
-    current_line: Line = Factory(Line)
+    current_line: Line = field(default_factory=Line)
     remove_u_prefix: bool = False
 
     def line(self, indent: int = 0) -> Iterator[Line]:
@@ -1716,46 +1958,13 @@ class LineGenerator(Visitor[Line]):
                 self.current_line.append(node)
         yield from super().visit_default(node)
 
-    def visit_atom(self, node: Node) -> Iterator[Line]:
-        # Always make parentheses invisible around a single node, because it should
-        # not be needed (except in the case of yield, where removing the parentheses
-        # produces a SyntaxError).
-        if (
-            len(node.children) == 3
-            and isinstance(node.children[0], Leaf)
-            and node.children[0].type == token.LPAR
-            and isinstance(node.children[2], Leaf)
-            and node.children[2].type == token.RPAR
-            and isinstance(node.children[1], Leaf)
-            and not (
-                node.children[1].type == token.NAME
-                and node.children[1].value == "yield"
-            )
-        ):
-            node.children[0].value = ""
-            node.children[2].value = ""
-        yield from super().visit_default(node)
-
-    def visit_factor(self, node: Node) -> Iterator[Line]:
-        """Force parentheses between a unary op and a binary power:
-
-        -2 ** 8 -> -(2 ** 8)
-        """
-        child = node.children[1]
-        if child.type == syms.power and len(child.children) == 3:
-            lpar = Leaf(token.LPAR, "(")
-            rpar = Leaf(token.RPAR, ")")
-            index = child.remove() or 0
-            node.insert_child(index, Node(syms.atom, [lpar, child, rpar]))
-        yield from self.visit_default(node)
-
-    def visit_INDENT(self, node: Node) -> Iterator[Line]:
+    def visit_INDENT(self, node: Leaf) -> Iterator[Line]:
         """Increase indentation level, maybe yield a line."""
         # In blib2to3 INDENT never holds comments.
         yield from self.line(+1)
         yield from self.visit_default(node)
 
-    def visit_DEDENT(self, node: Node) -> Iterator[Line]:
+    def visit_DEDENT(self, node: Leaf) -> Iterator[Line]:
         """Decrease indentation level, maybe yield a line."""
         # The current line might still wait for trailing comments.  At DEDENT time
         # there won't be any (they would be prefixes on the preceding NEWLINE).
@@ -1848,7 +2057,36 @@ class LineGenerator(Visitor[Line]):
             yield from self.line()
         yield from self.visit_default(leaf)
 
-    def __attrs_post_init__(self) -> None:
+    def visit_factor(self, node: Node) -> Iterator[Line]:
+        """Force parentheses between a unary op and a binary power:
+
+        -2 ** 8 -> -(2 ** 8)
+        """
+        _operator, operand = node.children
+        if (
+            operand.type == syms.power
+            and len(operand.children) == 3
+            and operand.children[1].type == token.DOUBLESTAR
+        ):
+            lpar = Leaf(token.LPAR, "(")
+            rpar = Leaf(token.RPAR, ")")
+            index = operand.remove() or 0
+            node.insert_child(index, Node(syms.atom, [lpar, operand, rpar]))
+        yield from self.visit_default(node)
+
+    def visit_STRING(self, leaf: Leaf) -> Iterator[Line]:
+        # Check if it's a docstring
+        if prev_siblings_are(
+            leaf.parent, [None, token.NEWLINE, token.INDENT, syms.simple_stmt]
+        ) and is_multiline_string(leaf):
+            prefix = "    " * self.current_line.depth
+            docstring = fix_docstring(leaf.value[3:-3], prefix)
+            leaf.value = leaf.value[0:3] + docstring + leaf.value[-3:]
+            normalize_string_quotes(leaf)
+
+        yield from self.visit_default(leaf)
+
+    def __post_init__(self) -> None:
         """You are in a twisty little maze of passages."""
         v = self.visit_stmt
         Ø: Set[str] = set()
@@ -2129,6 +2367,22 @@ def preceding_leaf(node: Optional[LN]) -> Optional[Leaf]:
     return None
 
 
+def prev_siblings_are(node: Optional[LN], tokens: List[Optional[NodeType]]) -> bool:
+    """Return if the `node` and its previous siblings match types against the provided
+    list of tokens; the provided `node`has its type matched against the last element in
+    the list.  `None` can be used as the first element to declare that the start of the
+    list is anchored at the start of its parent's children."""
+    if not tokens:
+        return True
+    if tokens[-1] is None:
+        return node is None
+    if not node:
+        return False
+    if node.type != tokens[-1]:
+        return False
+    return prev_siblings_are(node.prev_sibling, tokens[:-1])
+
+
 def child_towards(ancestor: Node, descendant: LN) -> Optional[LN]:
     """Return the child of `ancestor` that contains `descendant`."""
     node: Optional[LN] = descendant
@@ -2373,18 +2627,15 @@ def make_comment(content: str) -> str:
     return "#" + content
 
 
-def split_line(
+def transform_line(
     line: Line,
     line_length: int,
-    inner: bool = False,
+    normalize_strings: bool,
     features: Collection[Feature] = (),
 ) -> Iterator[Line]:
-    """Split a `line` into potentially many lines.
+    """Transform a `line`, potentially splitting it into many lines.
 
     They should fit in the allotted `line_length` but might not be able to.
-    `inner` signifies that there were a pair of brackets somewhere around the
-    current `line`, possibly transitively. This means we can fallback to splitting
-    by delimiters if the LHS/RHS don't yield any results.
 
     `features` are syntactical features that may be used in the output.
     """
@@ -2392,8 +2643,18 @@ def split_line(
         yield line
         return
 
-    line_str = str(line).strip("\n")
+    line_str = line_to_string(line)
+
+    def init_st(ST: Type[StringTransformer]) -> StringTransformer:
+        """Initialize StringTransformer"""
+        return ST(line_length, normalize_strings)
 
+    string_merge = init_st(StringMerger)
+    string_paren_strip = init_st(StringParenStripper)
+    string_split = init_st(StringSplitter)
+    string_paren_wrap = init_st(StringParenWrapper)
+
+    transformers: List[Transformer]
     if (
         not line.contains_uncollapsable_type_comments()
         and not line.should_explode
@@ -2402,13 +2663,12 @@ def split_line(
             is_line_short_enough(line, line_length=line_length, line_str=line_str)
             or line.contains_unsplittable_type_ignore()
         )
+        and not (line.contains_standalone_comments() and line.inside_brackets)
     ):
-        yield line
-        return
-
-    split_funcs: List[SplitFunc]
-    if line.is_def:
-        split_funcs = [left_hand_split]
+        # Only apply basic string preprocessing, since lines shouldn't be split here.
+        transformers = [string_merge, string_paren_strip]
+    elif line.is_def:
+        transformers = [left_hand_split]
     else:
 
         def rhs(line: Line, features: Collection[Feature]) -> Iterator[Line]:
@@ -2421,30 +2681,51 @@ def split_line(
             # All splits failed, best effort split with no omits.
             # This mostly happens to multiline strings that are by definition
             # reported as not fitting a single line.
-            yield from right_hand_split(line, line_length, features=features)
+            # line_length=1 here was historically a bug that somehow became a feature.
+            # See #762 and #781 for the full story.
+            yield from right_hand_split(line, line_length=1, features=features)
 
         if line.inside_brackets:
-            split_funcs = [delimiter_split, standalone_comment_split, rhs]
+            transformers = [
+                string_merge,
+                string_paren_strip,
+                delimiter_split,
+                standalone_comment_split,
+                string_split,
+                string_paren_wrap,
+                rhs,
+            ]
         else:
-            split_funcs = [rhs]
-    for split_func in split_funcs:
+            transformers = [
+                string_merge,
+                string_paren_strip,
+                string_split,
+                string_paren_wrap,
+                rhs,
+            ]
+
+    for transform in transformers:
         # We are accumulating lines in `result` because we might want to abort
         # mission and return the original line in the end, or attempt a different
         # split altogether.
         result: List[Line] = []
         try:
-            for l in split_func(line, features):
+            for l in transform(line, features):
                 if str(l).strip("\n") == line_str:
-                    raise CannotSplit("Split function returned an unchanged result")
+                    raise CannotTransform(
+                        "Line transformer returned an unchanged result"
+                    )
 
                 result.extend(
-                    split_line(
-                        l, line_length=line_length, inner=True, features=features
+                    transform_line(
+                        l,
+                        line_length=line_length,
+                        normalize_strings=normalize_strings,
+                        features=features,
                     )
                 )
-        except CannotSplit:
+        except CannotTransform:
             continue
-
         else:
             yield from result
             break
@@ -2453,70 +2734,2001 @@ def split_line(
         yield line
 
 
-def left_hand_split(line: Line, features: Collection[Feature] = ()) -> Iterator[Line]:
-    """Split line into many lines, starting with the first matching bracket pair.
-
-    Note: this usually looks weird, only use this for function definitions.
-    Prefer RHS otherwise.  This is why this function is not symmetrical with
-    :func:`right_hand_split` which also handles optional parentheses.
+@dataclass  # type: ignore
+class StringTransformer(ABC):
     """
-    tail_leaves: List[Leaf] = []
-    body_leaves: List[Leaf] = []
-    head_leaves: List[Leaf] = []
-    current_leaves = head_leaves
-    matching_bracket = None
-    for leaf in line.leaves:
-        if (
-            current_leaves is body_leaves
-            and leaf.type in CLOSING_BRACKETS
-            and leaf.opening_bracket is matching_bracket
-        ):
-            current_leaves = tail_leaves if body_leaves else head_leaves
-        current_leaves.append(leaf)
-        if current_leaves is head_leaves:
-            if leaf.type in OPENING_BRACKETS:
-                matching_bracket = leaf
-                current_leaves = body_leaves
-    if not matching_bracket:
-        raise CannotSplit("No brackets found")
+    An implementation of the Transformer protocol that relies on its
+    subclasses overriding the template methods `do_match(...)` and
+    `do_transform(...)`.
 
-    head = bracket_split_build_line(head_leaves, line, matching_bracket)
-    body = bracket_split_build_line(body_leaves, line, matching_bracket, is_body=True)
-    tail = bracket_split_build_line(tail_leaves, line, matching_bracket)
-    bracket_split_succeeded_or_raise(head, body, tail)
-    for result in (head, body, tail):
-        if result:
-            yield result
+    This Transformer works exclusively on strings (for example, by merging
+    or splitting them).
 
+    The following sections can be found among the docstrings of each concrete
+    StringTransformer subclass.
 
-def right_hand_split(
-    line: Line,
-    line_length: int,
-    features: Collection[Feature] = (),
-    omit: Collection[LeafID] = (),
-) -> Iterator[Line]:
-    """Split line into many lines, starting with the last matching bracket pair.
+    Requirements:
+        Which requirements must be met of the given Line for this
+        StringTransformer to be applied?
 
-    If the split was by optional parentheses, attempt splitting without them, too.
-    `omit` is a collection of closing bracket IDs that shouldn't be considered for
-    this split.
+    Transformations:
+        If the given Line meets all of the above requirments, which string
+        transformations can you expect to be applied to it by this
+        StringTransformer?
 
-    Note: running this function modifies `bracket_depth` on the leaves of `line`.
+    Collaborations:
+        What contractual agreements does this StringTransformer have with other
+        StringTransfomers? Such collaborations should be eliminated/minimized
+        as much as possible.
     """
-    tail_leaves: List[Leaf] = []
-    body_leaves: List[Leaf] = []
-    head_leaves: List[Leaf] = []
-    current_leaves = tail_leaves
-    opening_bracket = None
-    closing_bracket = None
-    for leaf in reversed(line.leaves):
-        if current_leaves is body_leaves:
-            if leaf is opening_bracket:
-                current_leaves = head_leaves if body_leaves else tail_leaves
-        current_leaves.append(leaf)
-        if current_leaves is tail_leaves:
-            if leaf.type in CLOSING_BRACKETS and id(leaf) not in omit:
-                opening_bracket = leaf.opening_bracket
+
+    line_length: int
+    normalize_strings: bool
+
+    @abstractmethod
+    def do_match(self, line: Line) -> TMatchResult:
+        """
+        Returns:
+            * Ok(string_idx) such that `line.leaves[string_idx]` is our target
+            string, if a match was able to be made.
+                OR
+            * Err(CannotTransform), if a match was not able to be made.
+        """
+
+    @abstractmethod
+    def do_transform(self, line: Line, string_idx: int) -> Iterator[TResult[Line]]:
+        """
+        Yields:
+            * Ok(new_line) where new_line is the new transformed line.
+                OR
+            * Err(CannotTransform) if the transformation failed for some reason. The
+            `do_match(...)` template method should usually be used to reject
+            the form of the given Line, but in some cases it is difficult to
+            know whether or not a Line meets the StringTransformer's
+            requirements until the transformation is already midway.
+
+        Side Effects:
+            This method should NOT mutate @line directly, but it MAY mutate the
+            Line's underlying Node structure. (WARNING: If the underlying Node
+            structure IS altered, then this method should NOT be allowed to
+            yield an CannotTransform after that point.)
+        """
+
+    def __call__(self, line: Line, _features: Collection[Feature]) -> Iterator[Line]:
+        """
+        StringTransformer instances have a call signature that mirrors that of
+        the Transformer type.
+
+        Raises:
+            CannotTransform(...) if the concrete StringTransformer class is unable
+            to transform @line.
+        """
+        # Optimization to avoid calling `self.do_match(...)` when the line does
+        # not contain any string.
+        if not any(leaf.type == token.STRING for leaf in line.leaves):
+            raise CannotTransform("There are no strings in this line.")
+
+        match_result = self.do_match(line)
+
+        if isinstance(match_result, Err):
+            cant_transform = match_result.err()
+            raise CannotTransform(
+                f"The string transformer {self.__class__.__name__} does not recognize"
+                " this line as one that it can transform."
+            ) from cant_transform
+
+        string_idx = match_result.ok()
+
+        for line_result in self.do_transform(line, string_idx):
+            if isinstance(line_result, Err):
+                cant_transform = line_result.err()
+                raise CannotTransform(
+                    "StringTransformer failed while attempting to transform string."
+                ) from cant_transform
+            line = line_result.ok()
+            yield line
+
+
+@dataclass
+class CustomSplit:
+    """A custom (i.e. manual) string split.
+
+    A single CustomSplit instance represents a single substring.
+
+    Examples:
+        Consider the following string:
+        ```
+        "Hi there friend."
+        " This is a custom"
+        f" string {split}."
+        ```
+
+        This string will correspond to the following three CustomSplit instances:
+        ```
+        CustomSplit(False, 16)
+        CustomSplit(False, 17)
+        CustomSplit(True, 16)
+        ```
+    """
+
+    has_prefix: bool
+    break_idx: int
+
+
+class CustomSplitMapMixin:
+    """
+    This mixin class is used to map merged strings to a sequence of
+    CustomSplits, which will then be used to re-split the strings iff none of
+    the resultant substrings go over the configured max line length.
+    """
+
+    _Key = Tuple[StringID, str]
+    _CUSTOM_SPLIT_MAP: Dict[_Key, Tuple[CustomSplit, ...]] = defaultdict(tuple)
+
+    @staticmethod
+    def _get_key(string: str) -> "CustomSplitMapMixin._Key":
+        """
+        Returns:
+            A unique identifier that is used internally to map @string to a
+            group of custom splits.
+        """
+        return (id(string), string)
+
+    def add_custom_splits(
+        self, string: str, custom_splits: Iterable[CustomSplit]
+    ) -> None:
+        """Custom Split Map Setter Method
+
+        Side Effects:
+            Adds a mapping from @string to the custom splits @custom_splits.
+        """
+        key = self._get_key(string)
+        self._CUSTOM_SPLIT_MAP[key] = tuple(custom_splits)
+
+    def pop_custom_splits(self, string: str) -> List[CustomSplit]:
+        """Custom Split Map Getter Method
+
+        Returns:
+            * A list of the custom splits that are mapped to @string, if any
+            exist.
+                OR
+            * [], otherwise.
+
+        Side Effects:
+            Deletes the mapping between @string and its associated custom
+            splits (which are returned to the caller).
+        """
+        key = self._get_key(string)
+
+        custom_splits = self._CUSTOM_SPLIT_MAP[key]
+        del self._CUSTOM_SPLIT_MAP[key]
+
+        return list(custom_splits)
+
+    def has_custom_splits(self, string: str) -> bool:
+        """
+        Returns:
+            True iff @string is associated with a set of custom splits.
+        """
+        key = self._get_key(string)
+        return key in self._CUSTOM_SPLIT_MAP
+
+
+class StringMerger(CustomSplitMapMixin, StringTransformer):
+    """StringTransformer that merges strings together.
+
+    Requirements:
+        (A) The line contains adjacent strings such that at most one substring
+        has inline comments AND none of those inline comments are pragmas AND
+        the set of all substring prefixes is either of length 1 or equal to
+        {"", "f"} AND none of the substrings are raw strings (i.e. are prefixed
+        with 'r').
+            OR
+        (B) The line contains a string which uses line continuation backslashes.
+
+    Transformations:
+        Depending on which of the two requirements above where met, either:
+
+        (A) The string group associated with the target string is merged.
+            OR
+        (B) All line-continuation backslashes are removed from the target string.
+
+    Collaborations:
+        StringMerger provides custom split information to StringSplitter.
+    """
+
+    def do_match(self, line: Line) -> TMatchResult:
+        LL = line.leaves
+
+        is_valid_index = is_valid_index_factory(LL)
+
+        for (i, leaf) in enumerate(LL):
+            if (
+                leaf.type == token.STRING
+                and is_valid_index(i + 1)
+                and LL[i + 1].type == token.STRING
+            ):
+                return Ok(i)
+
+            if leaf.type == token.STRING and "\\\n" in leaf.value:
+                return Ok(i)
+
+        return TErr("This line has no strings that need merging.")
+
+    def do_transform(self, line: Line, string_idx: int) -> Iterator[TResult[Line]]:
+        new_line = line
+        rblc_result = self.__remove_backslash_line_continuation_chars(
+            new_line, string_idx
+        )
+        if isinstance(rblc_result, Ok):
+            new_line = rblc_result.ok()
+
+        msg_result = self.__merge_string_group(new_line, string_idx)
+        if isinstance(msg_result, Ok):
+            new_line = msg_result.ok()
+
+        if isinstance(rblc_result, Err) and isinstance(msg_result, Err):
+            msg_cant_transform = msg_result.err()
+            rblc_cant_transform = rblc_result.err()
+            cant_transform = CannotTransform(
+                "StringMerger failed to merge any strings in this line."
+            )
+
+            # Chain the errors together using `__cause__`.
+            msg_cant_transform.__cause__ = rblc_cant_transform
+            cant_transform.__cause__ = msg_cant_transform
+
+            yield Err(cant_transform)
+        else:
+            yield Ok(new_line)
+
+    @staticmethod
+    def __remove_backslash_line_continuation_chars(
+        line: Line, string_idx: int
+    ) -> TResult[Line]:
+        """
+        Merge strings that were split across multiple lines using
+        line-continuation backslashes.
+
+        Returns:
+            Ok(new_line), if @line contains backslash line-continuation
+            characters.
+                OR
+            Err(CannotTransform), otherwise.
+        """
+        LL = line.leaves
+
+        string_leaf = LL[string_idx]
+        if not (
+            string_leaf.type == token.STRING
+            and "\\\n" in string_leaf.value
+            and not has_triple_quotes(string_leaf.value)
+        ):
+            return TErr(
+                f"String leaf {string_leaf} does not contain any backslash line"
+                " continuation characters."
+            )
+
+        new_line = line.clone()
+        new_line.comments = line.comments
+        append_leaves(new_line, line, LL)
+
+        new_string_leaf = new_line.leaves[string_idx]
+        new_string_leaf.value = new_string_leaf.value.replace("\\\n", "")
+
+        return Ok(new_line)
+
+    def __merge_string_group(self, line: Line, string_idx: int) -> TResult[Line]:
+        """
+        Merges string group (i.e. set of adjacent strings) where the first
+        string in the group is `line.leaves[string_idx]`.
+
+        Returns:
+            Ok(new_line), if ALL of the validation checks found in
+            __validate_msg(...) pass.
+                OR
+            Err(CannotTransform), otherwise.
+        """
+        LL = line.leaves
+
+        is_valid_index = is_valid_index_factory(LL)
+
+        vresult = self.__validate_msg(line, string_idx)
+        if isinstance(vresult, Err):
+            return vresult
+
+        # If the string group is wrapped inside an Atom node, we must make sure
+        # to later replace that Atom with our new (merged) string leaf.
+        atom_node = LL[string_idx].parent
+
+        # We will place BREAK_MARK in between every two substrings that we
+        # merge. We will then later go through our final result and use the
+        # various instances of BREAK_MARK we find to add the right values to
+        # the custom split map.
+        BREAK_MARK = "@@@@@ BLACK BREAKPOINT MARKER @@@@@"
+
+        QUOTE = LL[string_idx].value[-1]
+
+        def make_naked(string: str, string_prefix: str) -> str:
+            """Strip @string (i.e. make it a "naked" string)
+
+            Pre-conditions:
+                * assert_is_leaf_string(@string)
+
+            Returns:
+                A string that is identical to @string except that
+                @string_prefix has been stripped, the surrounding QUOTE
+                characters have been removed, and any remaining QUOTE
+                characters have been escaped.
+            """
+            assert_is_leaf_string(string)
+
+            RE_EVEN_BACKSLASHES = r"(?:(?<!\\)(?:\\\\)*)"
+            naked_string = string[len(string_prefix) + 1 : -1]
+            naked_string = re.sub(
+                "(" + RE_EVEN_BACKSLASHES + ")" + QUOTE, r"\1\\" + QUOTE, naked_string
+            )
+            return naked_string
+
+        # Holds the CustomSplit objects that will later be added to the custom
+        # split map.
+        custom_splits = []
+
+        # Temporary storage for the 'has_prefix' part of the CustomSplit objects.
+        prefix_tracker = []
+
+        # Sets the 'prefix' variable. This is the prefix that the final merged
+        # string will have.
+        next_str_idx = string_idx
+        prefix = ""
+        while (
+            not prefix
+            and is_valid_index(next_str_idx)
+            and LL[next_str_idx].type == token.STRING
+        ):
+            prefix = get_string_prefix(LL[next_str_idx].value)
+            next_str_idx += 1
+
+        # The next loop merges the string group. The final string will be
+        # contained in 'S'.
+        #
+        # The following convenience variables are used:
+        #
+        #   S: string
+        #   NS: naked string
+        #   SS: next string
+        #   NSS: naked next string
+        S = ""
+        NS = ""
+        num_of_strings = 0
+        next_str_idx = string_idx
+        while is_valid_index(next_str_idx) and LL[next_str_idx].type == token.STRING:
+            num_of_strings += 1
+
+            SS = LL[next_str_idx].value
+            next_prefix = get_string_prefix(SS)
+
+            # If this is an f-string group but this substring is not prefixed
+            # with 'f'...
+            if "f" in prefix and "f" not in next_prefix:
+                # Then we must escape any braces contained in this substring.
+                SS = re.subf(r"(\{|\})", "{1}{1}", SS)
+
+            NSS = make_naked(SS, next_prefix)
+
+            has_prefix = bool(next_prefix)
+            prefix_tracker.append(has_prefix)
+
+            S = prefix + QUOTE + NS + NSS + BREAK_MARK + QUOTE
+            NS = make_naked(S, prefix)
+
+            next_str_idx += 1
+
+        S_leaf = Leaf(token.STRING, S)
+        if self.normalize_strings:
+            normalize_string_quotes(S_leaf)
+
+        # Fill the 'custom_splits' list with the appropriate CustomSplit objects.
+        temp_string = S_leaf.value[len(prefix) + 1 : -1]
+        for has_prefix in prefix_tracker:
+            mark_idx = temp_string.find(BREAK_MARK)
+            assert (
+                mark_idx >= 0
+            ), "Logic error while filling the custom string breakpoint cache."
+
+            temp_string = temp_string[mark_idx + len(BREAK_MARK) :]
+            breakpoint_idx = mark_idx + (len(prefix) if has_prefix else 0) + 1
+            custom_splits.append(CustomSplit(has_prefix, breakpoint_idx))
+
+        string_leaf = Leaf(token.STRING, S_leaf.value.replace(BREAK_MARK, ""))
+
+        if atom_node is not None:
+            replace_child(atom_node, string_leaf)
+
+        # Build the final line ('new_line') that this method will later return.
+        new_line = line.clone()
+        for (i, leaf) in enumerate(LL):
+            if i == string_idx:
+                new_line.append(string_leaf)
+
+            if string_idx <= i < string_idx + num_of_strings:
+                for comment_leaf in line.comments_after(LL[i]):
+                    new_line.append(comment_leaf, preformatted=True)
+                continue
+
+            append_leaves(new_line, line, [leaf])
+
+        self.add_custom_splits(string_leaf.value, custom_splits)
+        return Ok(new_line)
+
+    @staticmethod
+    def __validate_msg(line: Line, string_idx: int) -> TResult[None]:
+        """Validate (M)erge (S)tring (G)roup
+
+        Transform-time string validation logic for __merge_string_group(...).
+
+        Returns:
+            * Ok(None), if ALL validation checks (listed below) pass.
+                OR
+            * Err(CannotTransform), if any of the following are true:
+                - The target string is not in a string group (i.e. it has no
+                  adjacent strings).
+                - The string group has more than one inline comment.
+                - The string group has an inline comment that appears to be a pragma.
+                - The set of all string prefixes in the string group is of
+                  length greater than one and is not equal to {"", "f"}.
+                - The string group consists of raw strings.
+        """
+        num_of_inline_string_comments = 0
+        set_of_prefixes = set()
+        num_of_strings = 0
+        for leaf in line.leaves[string_idx:]:
+            if leaf.type != token.STRING:
+                # If the string group is trailed by a comma, we count the
+                # comments trailing the comma to be one of the string group's
+                # comments.
+                if leaf.type == token.COMMA and id(leaf) in line.comments:
+                    num_of_inline_string_comments += 1
+                break
+
+            if has_triple_quotes(leaf.value):
+                return TErr("StringMerger does NOT merge multiline strings.")
+
+            num_of_strings += 1
+            prefix = get_string_prefix(leaf.value)
+            if "r" in prefix:
+                return TErr("StringMerger does NOT merge raw strings.")
+
+            set_of_prefixes.add(prefix)
+
+            if id(leaf) in line.comments:
+                num_of_inline_string_comments += 1
+                if contains_pragma_comment(line.comments[id(leaf)]):
+                    return TErr("Cannot merge strings which have pragma comments.")
+
+        if num_of_strings < 2:
+            return TErr(
+                f"Not enough strings to merge (num_of_strings={num_of_strings})."
+            )
+
+        if num_of_inline_string_comments > 1:
+            return TErr(
+                f"Too many inline string comments ({num_of_inline_string_comments})."
+            )
+
+        if len(set_of_prefixes) > 1 and set_of_prefixes != {"", "f"}:
+            return TErr(f"Too many different prefixes ({set_of_prefixes}).")
+
+        return Ok(None)
+
+
+class StringParenStripper(StringTransformer):
+    """StringTransformer that strips surrounding parentheses from strings.
+
+    Requirements:
+        The line contains a string which is surrounded by parentheses and:
+            - The target string is NOT the only argument to a function call).
+            - The RPAR is NOT followed by an attribute access (i.e. a dot).
+
+    Transformations:
+        The parentheses mentioned in the 'Requirements' section are stripped.
+
+    Collaborations:
+        StringParenStripper has its own inherent usefulness, but it is also
+        relied on to clean up the parentheses created by StringParenWrapper (in
+        the event that they are no longer needed).
+    """
+
+    def do_match(self, line: Line) -> TMatchResult:
+        LL = line.leaves
+
+        is_valid_index = is_valid_index_factory(LL)
+
+        for (idx, leaf) in enumerate(LL):
+            # Should be a string...
+            if leaf.type != token.STRING:
+                continue
+
+            # Should be preceded by a non-empty LPAR...
+            if (
+                not is_valid_index(idx - 1)
+                or LL[idx - 1].type != token.LPAR
+                or is_empty_lpar(LL[idx - 1])
+            ):
+                continue
+
+            # That LPAR should NOT be preceded by a function name or a closing
+            # bracket (which could be a function which returns a function or a
+            # list/dictionary that contains a function)...
+            if is_valid_index(idx - 2) and (
+                LL[idx - 2].type == token.NAME or LL[idx - 2].type in CLOSING_BRACKETS
+            ):
+                continue
+
+            string_idx = idx
+
+            # Skip the string trailer, if one exists.
+            string_parser = StringParser()
+            next_idx = string_parser.parse(LL, string_idx)
+
+            # Should be followed by a non-empty RPAR...
+            if (
+                is_valid_index(next_idx)
+                and LL[next_idx].type == token.RPAR
+                and not is_empty_rpar(LL[next_idx])
+            ):
+                # That RPAR should NOT be followed by a '.' symbol.
+                if is_valid_index(next_idx + 1) and LL[next_idx + 1].type == token.DOT:
+                    continue
+
+                return Ok(string_idx)
+
+        return TErr("This line has no strings wrapped in parens.")
+
+    def do_transform(self, line: Line, string_idx: int) -> Iterator[TResult[Line]]:
+        LL = line.leaves
+
+        string_parser = StringParser()
+        rpar_idx = string_parser.parse(LL, string_idx)
+
+        for leaf in (LL[string_idx - 1], LL[rpar_idx]):
+            if line.comments_after(leaf):
+                yield TErr(
+                    "Will not strip parentheses which have comments attached to them."
+                )
+
+        new_line = line.clone()
+        new_line.comments = line.comments.copy()
+
+        append_leaves(new_line, line, LL[: string_idx - 1])
+
+        string_leaf = Leaf(token.STRING, LL[string_idx].value)
+        LL[string_idx - 1].remove()
+        replace_child(LL[string_idx], string_leaf)
+        new_line.append(string_leaf)
+
+        append_leaves(
+            new_line, line, LL[string_idx + 1 : rpar_idx] + LL[rpar_idx + 1 :],
+        )
+
+        LL[rpar_idx].remove()
+
+        yield Ok(new_line)
+
+
+class BaseStringSplitter(StringTransformer):
+    """
+    Abstract class for StringTransformers which transform a Line's strings by splitting
+    them or placing them on their own lines where necessary to avoid going over
+    the configured line length.
+
+    Requirements:
+        * The target string value is responsible for the line going over the
+        line length limit. It follows that after all of black's other line
+        split methods have been exhausted, this line (or one of the resulting
+        lines after all line splits are performed) would still be over the
+        line_length limit unless we split this string.
+            AND
+        * The target string is NOT a "pointless" string (i.e. a string that has
+        no parent or siblings).
+            AND
+        * The target string is not followed by an inline comment that appears
+        to be a pragma.
+            AND
+        * The target string is not a multiline (i.e. triple-quote) string.
+    """
+
+    @abstractmethod
+    def do_splitter_match(self, line: Line) -> TMatchResult:
+        """
+        BaseStringSplitter asks its clients to override this method instead of
+        `StringTransformer.do_match(...)`.
+
+        Follows the same protocol as `StringTransformer.do_match(...)`.
+
+        Refer to `help(StringTransformer.do_match)` for more information.
+        """
+
+    def do_match(self, line: Line) -> TMatchResult:
+        match_result = self.do_splitter_match(line)
+        if isinstance(match_result, Err):
+            return match_result
+
+        string_idx = match_result.ok()
+        vresult = self.__validate(line, string_idx)
+        if isinstance(vresult, Err):
+            return vresult
+
+        return match_result
+
+    def __validate(self, line: Line, string_idx: int) -> TResult[None]:
+        """
+        Checks that @line meets all of the requirements listed in this classes'
+        docstring. Refer to `help(BaseStringSplitter)` for a detailed
+        description of those requirements.
+
+        Returns:
+            * Ok(None), if ALL of the requirements are met.
+                OR
+            * Err(CannotTransform), if ANY of the requirements are NOT met.
+        """
+        LL = line.leaves
+
+        string_leaf = LL[string_idx]
+
+        max_string_length = self.__get_max_string_length(line, string_idx)
+        if len(string_leaf.value) <= max_string_length:
+            return TErr(
+                "The string itself is not what is causing this line to be too long."
+            )
+
+        if not string_leaf.parent or [L.type for L in string_leaf.parent.children] == [
+            token.STRING,
+            token.NEWLINE,
+        ]:
+            return TErr(
+                f"This string ({string_leaf.value}) appears to be pointless (i.e. has"
+                " no parent)."
+            )
+
+        if id(line.leaves[string_idx]) in line.comments and contains_pragma_comment(
+            line.comments[id(line.leaves[string_idx])]
+        ):
+            return TErr(
+                "Line appears to end with an inline pragma comment. Splitting the line"
+                " could modify the pragma's behavior."
+            )
+
+        if has_triple_quotes(string_leaf.value):
+            return TErr("We cannot split multiline strings.")
+
+        return Ok(None)
+
+    def __get_max_string_length(self, line: Line, string_idx: int) -> int:
+        """
+        Calculates the max string length used when attempting to determine
+        whether or not the target string is responsible for causing the line to
+        go over the line length limit.
+
+        WARNING: This method is tightly coupled to both StringSplitter and
+        (especially) StringParenWrapper. There is probably a better way to
+        accomplish what is being done here.
+
+        Returns:
+            max_string_length: such that `line.leaves[string_idx].value >
+            max_string_length` implies that the target string IS responsible
+            for causing this line to exceed the line length limit.
+        """
+        LL = line.leaves
+
+        is_valid_index = is_valid_index_factory(LL)
+
+        # We use the shorthand "WMA4" in comments to abbreviate "We must
+        # account for". When giving examples, we use STRING to mean some/any
+        # valid string.
+        #
+        # Finally, we use the following convenience variables:
+        #
+        #   P:  The leaf that is before the target string leaf.
+        #   N:  The leaf that is after the target string leaf.
+        #   NN: The leaf that is after N.
+
+        # WMA4 the whitespace at the beginning of the line.
+        offset = line.depth * 4
+
+        if is_valid_index(string_idx - 1):
+            p_idx = string_idx - 1
+            if (
+                LL[string_idx - 1].type == token.LPAR
+                and LL[string_idx - 1].value == ""
+                and string_idx >= 2
+            ):
+                # If the previous leaf is an empty LPAR placeholder, we should skip it.
+                p_idx -= 1
+
+            P = LL[p_idx]
+            if P.type == token.PLUS:
+                # WMA4 a space and a '+' character (e.g. `+ STRING`).
+                offset += 2
+
+            if P.type == token.COMMA:
+                # WMA4 a space, a comma, and a closing bracket [e.g. `), STRING`].
+                offset += 3
+
+            if P.type in [token.COLON, token.EQUAL, token.NAME]:
+                # This conditional branch is meant to handle dictionary keys,
+                # variable assignments, 'return STRING' statement lines, and
+                # 'else STRING' ternary expression lines.
+
+                # WMA4 a single space.
+                offset += 1
+
+                # WMA4 the lengths of any leaves that came before that space.
+                for leaf in LL[: p_idx + 1]:
+                    offset += len(str(leaf))
+
+        if is_valid_index(string_idx + 1):
+            N = LL[string_idx + 1]
+            if N.type == token.RPAR and N.value == "" and len(LL) > string_idx + 2:
+                # If the next leaf is an empty RPAR placeholder, we should skip it.
+                N = LL[string_idx + 2]
+
+            if N.type == token.COMMA:
+                # WMA4 a single comma at the end of the string (e.g `STRING,`).
+                offset += 1
+
+            if is_valid_index(string_idx + 2):
+                NN = LL[string_idx + 2]
+
+                if N.type == token.DOT and NN.type == token.NAME:
+                    # This conditional branch is meant to handle method calls invoked
+                    # off of a string literal up to and including the LPAR character.
+
+                    # WMA4 the '.' character.
+                    offset += 1
+
+                    if (
+                        is_valid_index(string_idx + 3)
+                        and LL[string_idx + 3].type == token.LPAR
+                    ):
+                        # WMA4 the left parenthesis character.
+                        offset += 1
+
+                    # WMA4 the length of the method's name.
+                    offset += len(NN.value)
+
+        has_comments = False
+        for comment_leaf in line.comments_after(LL[string_idx]):
+            if not has_comments:
+                has_comments = True
+                # WMA4 two spaces before the '#' character.
+                offset += 2
+
+            # WMA4 the length of the inline comment.
+            offset += len(comment_leaf.value)
+
+        max_string_length = self.line_length - offset
+        return max_string_length
+
+
+class StringSplitter(CustomSplitMapMixin, BaseStringSplitter):
+    """
+    StringTransformer that splits "atom" strings (i.e. strings which exist on
+    lines by themselves).
+
+    Requirements:
+        * The line consists ONLY of a single string (with the exception of a
+        '+' symbol which MAY exist at the start of the line), MAYBE a string
+        trailer, and MAYBE a trailing comma.
+            AND
+        * All of the requirements listed in BaseStringSplitter's docstring.
+
+    Transformations:
+        The string mentioned in the 'Requirements' section is split into as
+        many substrings as necessary to adhere to the configured line length.
+
+        In the final set of substrings, no substring should be smaller than
+        MIN_SUBSTR_SIZE characters.
+
+        The string will ONLY be split on spaces (i.e. each new substring should
+        start with a space).
+
+        If the string is an f-string, it will NOT be split in the middle of an
+        f-expression (e.g. in f"FooBar: {foo() if x else bar()}", {foo() if x
+        else bar()} is an f-expression).
+
+        If the string that is being split has an associated set of custom split
+        records and those custom splits will NOT result in any line going over
+        the configured line length, those custom splits are used. Otherwise the
+        string is split as late as possible (from left-to-right) while still
+        adhering to the transformation rules listed above.
+
+    Collaborations:
+        StringSplitter relies on StringMerger to construct the appropriate
+        CustomSplit objects and add them to the custom split map.
+    """
+
+    MIN_SUBSTR_SIZE = 6
+    # Matches an "f-expression" (e.g. {var}) that might be found in an f-string.
+    RE_FEXPR = r"""
+    (?<!\{)\{
+        (?:
+            [^\{\}]
+            | \{\{
+            | \}\}
+        )+?
+    (?<!\})(?:\}\})*\}(?!\})
+    """
+
+    def do_splitter_match(self, line: Line) -> TMatchResult:
+        LL = line.leaves
+
+        is_valid_index = is_valid_index_factory(LL)
+
+        idx = 0
+
+        # The first leaf MAY be a '+' symbol...
+        if is_valid_index(idx) and LL[idx].type == token.PLUS:
+            idx += 1
+
+        # The next/first leaf MAY be an empty LPAR...
+        if is_valid_index(idx) and is_empty_lpar(LL[idx]):
+            idx += 1
+
+        # The next/first leaf MUST be a string...
+        if not is_valid_index(idx) or LL[idx].type != token.STRING:
+            return TErr("Line does not start with a string.")
+
+        string_idx = idx
+
+        # Skip the string trailer, if one exists.
+        string_parser = StringParser()
+        idx = string_parser.parse(LL, string_idx)
+
+        # That string MAY be followed by an empty RPAR...
+        if is_valid_index(idx) and is_empty_rpar(LL[idx]):
+            idx += 1
+
+        # That string / empty RPAR leaf MAY be followed by a comma...
+        if is_valid_index(idx) and LL[idx].type == token.COMMA:
+            idx += 1
+
+        # But no more leaves are allowed...
+        if is_valid_index(idx):
+            return TErr("This line does not end with a string.")
+
+        return Ok(string_idx)
+
+    def do_transform(self, line: Line, string_idx: int) -> Iterator[TResult[Line]]:
+        LL = line.leaves
+
+        QUOTE = LL[string_idx].value[-1]
+
+        is_valid_index = is_valid_index_factory(LL)
+        insert_str_child = insert_str_child_factory(LL[string_idx])
+
+        prefix = get_string_prefix(LL[string_idx].value)
+
+        # We MAY choose to drop the 'f' prefix from substrings that don't
+        # contain any f-expressions, but ONLY if the original f-string
+        # containes at least one f-expression. Otherwise, we will alter the AST
+        # of the program.
+        drop_pointless_f_prefix = ("f" in prefix) and re.search(
+            self.RE_FEXPR, LL[string_idx].value, re.VERBOSE
+        )
+
+        first_string_line = True
+        starts_with_plus = LL[0].type == token.PLUS
+
+        def line_needs_plus() -> bool:
+            return first_string_line and starts_with_plus
+
+        def maybe_append_plus(new_line: Line) -> None:
+            """
+            Side Effects:
+                If @line starts with a plus and this is the first line we are
+                constructing, this function appends a PLUS leaf to @new_line
+                and replaces the old PLUS leaf in the node structure. Otherwise
+                this function does nothing.
+            """
+            if line_needs_plus():
+                plus_leaf = Leaf(token.PLUS, "+")
+                replace_child(LL[0], plus_leaf)
+                new_line.append(plus_leaf)
+
+        ends_with_comma = (
+            is_valid_index(string_idx + 1) and LL[string_idx + 1].type == token.COMMA
+        )
+
+        def max_last_string() -> int:
+            """
+            Returns:
+                The max allowed length of the string value used for the last
+                line we will construct.
+            """
+            result = self.line_length
+            result -= line.depth * 4
+            result -= 1 if ends_with_comma else 0
+            result -= 2 if line_needs_plus() else 0
+            return result
+
+        # --- Calculate Max Break Index (for string value)
+        # We start with the line length limit
+        max_break_idx = self.line_length
+        # The last index of a string of length N is N-1.
+        max_break_idx -= 1
+        # Leading whitespace is not present in the string value (e.g. Leaf.value).
+        max_break_idx -= line.depth * 4
+        if max_break_idx < 0:
+            yield TErr(
+                f"Unable to split {LL[string_idx].value} at such high of a line depth:"
+                f" {line.depth}"
+            )
+            return
+
+        # Check if StringMerger registered any custom splits.
+        custom_splits = self.pop_custom_splits(LL[string_idx].value)
+        # We use them ONLY if none of them would produce lines that exceed the
+        # line limit.
+        use_custom_breakpoints = bool(
+            custom_splits
+            and all(csplit.break_idx <= max_break_idx for csplit in custom_splits)
+        )
+
+        # Temporary storage for the remaining chunk of the string line that
+        # can't fit onto the line currently being constructed.
+        rest_value = LL[string_idx].value
+
+        def more_splits_should_be_made() -> bool:
+            """
+            Returns:
+                True iff `rest_value` (the remaining string value from the last
+                split), should be split again.
+            """
+            if use_custom_breakpoints:
+                return len(custom_splits) > 1
+            else:
+                return len(rest_value) > max_last_string()
+
+        string_line_results: List[Ok[Line]] = []
+        while more_splits_should_be_made():
+            if use_custom_breakpoints:
+                # Custom User Split (manual)
+                csplit = custom_splits.pop(0)
+                break_idx = csplit.break_idx
+            else:
+                # Algorithmic Split (automatic)
+                max_bidx = max_break_idx - 2 if line_needs_plus() else max_break_idx
+                maybe_break_idx = self.__get_break_idx(rest_value, max_bidx)
+                if maybe_break_idx is None:
+                    # If we are unable to algorthmically determine a good split
+                    # and this string has custom splits registered to it, we
+                    # fall back to using them--which means we have to start
+                    # over from the beginning.
+                    if custom_splits:
+                        rest_value = LL[string_idx].value
+                        string_line_results = []
+                        first_string_line = True
+                        use_custom_breakpoints = True
+                        continue
+
+                    # Otherwise, we stop splitting here.
+                    break
+
+                break_idx = maybe_break_idx
+
+            # --- Construct `next_value`
+            next_value = rest_value[:break_idx] + QUOTE
+            if (
+                # Are we allowed to try to drop a pointless 'f' prefix?
+                drop_pointless_f_prefix
+                # If we are, will we be successful?
+                and next_value != self.__normalize_f_string(next_value, prefix)
+            ):
+                # If the current custom split did NOT originally use a prefix,
+                # then `csplit.break_idx` will be off by one after removing
+                # the 'f' prefix.
+                break_idx = (
+                    break_idx + 1
+                    if use_custom_breakpoints and not csplit.has_prefix
+                    else break_idx
+                )
+                next_value = rest_value[:break_idx] + QUOTE
+                next_value = self.__normalize_f_string(next_value, prefix)
+
+            # --- Construct `next_leaf`
+            next_leaf = Leaf(token.STRING, next_value)
+            insert_str_child(next_leaf)
+            self.__maybe_normalize_string_quotes(next_leaf)
+
+            # --- Construct `next_line`
+            next_line = line.clone()
+            maybe_append_plus(next_line)
+            next_line.append(next_leaf)
+            string_line_results.append(Ok(next_line))
+
+            rest_value = prefix + QUOTE + rest_value[break_idx:]
+            first_string_line = False
+
+        yield from string_line_results
+
+        if drop_pointless_f_prefix:
+            rest_value = self.__normalize_f_string(rest_value, prefix)
+
+        rest_leaf = Leaf(token.STRING, rest_value)
+        insert_str_child(rest_leaf)
+
+        # NOTE: I could not find a test case that verifies that the following
+        # line is actually necessary, but it seems to be. Otherwise we risk
+        # not normalizing the last substring, right?
+        self.__maybe_normalize_string_quotes(rest_leaf)
+
+        last_line = line.clone()
+        maybe_append_plus(last_line)
+
+        # If there are any leaves to the right of the target string...
+        if is_valid_index(string_idx + 1):
+            # We use `temp_value` here to determine how long the last line
+            # would be if we were to append all the leaves to the right of the
+            # target string to the last string line.
+            temp_value = rest_value
+            for leaf in LL[string_idx + 1 :]:
+                temp_value += str(leaf)
+                if leaf.type == token.LPAR:
+                    break
+
+            # Try to fit them all on the same line with the last substring...
+            if (
+                len(temp_value) <= max_last_string()
+                or LL[string_idx + 1].type == token.COMMA
+            ):
+                last_line.append(rest_leaf)
+                append_leaves(last_line, line, LL[string_idx + 1 :])
+                yield Ok(last_line)
+            # Otherwise, place the last substring on one line and everything
+            # else on a line below that...
+            else:
+                last_line.append(rest_leaf)
+                yield Ok(last_line)
+
+                non_string_line = line.clone()
+                append_leaves(non_string_line, line, LL[string_idx + 1 :])
+                yield Ok(non_string_line)
+        # Else the target string was the last leaf...
+        else:
+            last_line.append(rest_leaf)
+            last_line.comments = line.comments.copy()
+            yield Ok(last_line)
+
+    def __get_break_idx(self, string: str, max_break_idx: int) -> Optional[int]:
+        """
+        This method contains the algorithm that StringSplitter uses to
+        determine which character to split each string at.
+
+        Args:
+            @string: The substring that we are attempting to split.
+            @max_break_idx: The ideal break index. We will return this value if it
+            meets all the necessary conditions. In the likely event that it
+            doesn't we will try to find the closest index BELOW @max_break_idx
+            that does. If that fails, we will expand our search by also
+            considering all valid indices ABOVE @max_break_idx.
+
+        Pre-Conditions:
+            * assert_is_leaf_string(@string)
+            * 0 <= @max_break_idx < len(@string)
+
+        Returns:
+            break_idx, if an index is able to be found that meets all of the
+            conditions listed in the 'Transformations' section of this classes'
+            docstring.
+                OR
+            None, otherwise.
+        """
+        is_valid_index = is_valid_index_factory(string)
+
+        assert is_valid_index(max_break_idx)
+        assert_is_leaf_string(string)
+
+        _fexpr_slices: Optional[List[Tuple[Index, Index]]] = None
+
+        def fexpr_slices() -> Iterator[Tuple[Index, Index]]:
+            """
+            Yields:
+                All ranges of @string which, if @string were to be split there,
+                would result in the splitting of an f-expression (which is NOT
+                allowed).
+            """
+            nonlocal _fexpr_slices
+
+            if _fexpr_slices is None:
+                _fexpr_slices = []
+                for match in re.finditer(self.RE_FEXPR, string, re.VERBOSE):
+                    _fexpr_slices.append(match.span())
+
+            yield from _fexpr_slices
+
+        is_fstring = "f" in get_string_prefix(string)
+
+        def breaks_fstring_expression(i: Index) -> bool:
+            """
+            Returns:
+                True iff returning @i would result in the splitting of an
+                f-expression (which is NOT allowed).
+            """
+            if not is_fstring:
+                return False
+
+            for (start, end) in fexpr_slices():
+                if start <= i < end:
+                    return True
+
+            return False
+
+        def passes_all_checks(i: Index) -> bool:
+            """
+            Returns:
+                True iff ALL of the conditions listed in the 'Transformations'
+                section of this classes' docstring would be be met by returning @i.
+            """
+            is_space = string[i] == " "
+            is_big_enough = (
+                len(string[i:]) >= self.MIN_SUBSTR_SIZE
+                and len(string[:i]) >= self.MIN_SUBSTR_SIZE
+            )
+            return is_space and is_big_enough and not breaks_fstring_expression(i)
+
+        # First, we check all indices BELOW @max_break_idx.
+        break_idx = max_break_idx
+        while is_valid_index(break_idx - 1) and not passes_all_checks(break_idx):
+            break_idx -= 1
+
+        if not passes_all_checks(break_idx):
+            # If that fails, we check all indices ABOVE @max_break_idx.
+            #
+            # If we are able to find a valid index here, the next line is going
+            # to be longer than the specified line length, but it's probably
+            # better than doing nothing at all.
+            break_idx = max_break_idx + 1
+            while is_valid_index(break_idx + 1) and not passes_all_checks(break_idx):
+                break_idx += 1
+
+            if not is_valid_index(break_idx) or not passes_all_checks(break_idx):
+                return None
+
+        return break_idx
+
+    def __maybe_normalize_string_quotes(self, leaf: Leaf) -> None:
+        if self.normalize_strings:
+            normalize_string_quotes(leaf)
+
+    def __normalize_f_string(self, string: str, prefix: str) -> str:
+        """
+        Pre-Conditions:
+            * assert_is_leaf_string(@string)
+
+        Returns:
+            * If @string is an f-string that contains no f-expressions, we
+            return a string identical to @string except that the 'f' prefix
+            has been stripped and all double braces (i.e. '{{' or '}}') have
+            been normalized (i.e. turned into '{' or '}').
+                OR
+            * Otherwise, we return @string.
+        """
+        assert_is_leaf_string(string)
+
+        if "f" in prefix and not re.search(self.RE_FEXPR, string, re.VERBOSE):
+            new_prefix = prefix.replace("f", "")
+
+            temp = string[len(prefix) :]
+            temp = re.sub(r"\{\{", "{", temp)
+            temp = re.sub(r"\}\}", "}", temp)
+            new_string = temp
+
+            return f"{new_prefix}{new_string}"
+        else:
+            return string
+
+
+class StringParenWrapper(CustomSplitMapMixin, BaseStringSplitter):
+    """
+    StringTransformer that splits non-"atom" strings (i.e. strings that do not
+    exist on lines by themselves).
+
+    Requirements:
+        All of the requirements listed in BaseStringSplitter's docstring in
+        addition to the requirements listed below:
+
+        * The line is a return/yield statement, which returns/yields a string.
+            OR
+        * The line is part of a ternary expression (e.g. `x = y if cond else
+        z`) such that the line starts with `else <string>`, where <string> is
+        some string.
+            OR
+        * The line is an assert statement, which ends with a string.
+            OR
+        * The line is an assignment statement (e.g. `x = <string>` or `x +=
+        <string>`) such that the variable is being assigned the value of some
+        string.
+            OR
+        * The line is a dictionary key assignment where some valid key is being
+        assigned the value of some string.
+
+    Transformations:
+        The chosen string is wrapped in parentheses and then split at the LPAR.
+
+        We then have one line which ends with an LPAR and another line that
+        starts with the chosen string. The latter line is then split again at
+        the RPAR. This results in the RPAR (and possibly a trailing comma)
+        being placed on its own line.
+
+        NOTE: If any leaves exist to the right of the chosen string (except
+        for a trailing comma, which would be placed after the RPAR), those
+        leaves are placed inside the parentheses.  In effect, the chosen
+        string is not necessarily being "wrapped" by parentheses. We can,
+        however, count on the LPAR being placed directly before the chosen
+        string.
+
+        In other words, StringParenWrapper creates "atom" strings. These
+        can then be split again by StringSplitter, if necessary.
+
+    Collaborations:
+        In the event that a string line split by StringParenWrapper is
+        changed such that it no longer needs to be given its own line,
+        StringParenWrapper relies on StringParenStripper to clean up the
+        parentheses it created.
+    """
+
+    def do_splitter_match(self, line: Line) -> TMatchResult:
+        LL = line.leaves
+
+        string_idx = None
+        string_idx = string_idx or self._return_match(LL)
+        string_idx = string_idx or self._else_match(LL)
+        string_idx = string_idx or self._assert_match(LL)
+        string_idx = string_idx or self._assign_match(LL)
+        string_idx = string_idx or self._dict_match(LL)
+
+        if string_idx is not None:
+            string_value = line.leaves[string_idx].value
+            # If the string has no spaces...
+            if " " not in string_value:
+                # And will still violate the line length limit when split...
+                max_string_length = self.line_length - ((line.depth + 1) * 4)
+                if len(string_value) > max_string_length:
+                    # And has no associated custom splits...
+                    if not self.has_custom_splits(string_value):
+                        # Then we should NOT put this string on its own line.
+                        return TErr(
+                            "We do not wrap long strings in parentheses when the"
+                            " resultant line would still be over the specified line"
+                            " length and can't be split further by StringSplitter."
+                        )
+            return Ok(string_idx)
+
+        return TErr("This line does not contain any non-atomic strings.")
+
+    @staticmethod
+    def _return_match(LL: List[Leaf]) -> Optional[int]:
+        """
+        Returns:
+            string_idx such that @LL[string_idx] is equal to our target (i.e.
+            matched) string, if this line matches the return/yield statement
+            requirements listed in the 'Requirements' section of this classes'
+            docstring.
+                OR
+            None, otherwise.
+        """
+        # If this line is apart of a return/yield statement and the first leaf
+        # contains either the "return" or "yield" keywords...
+        if parent_type(LL[0]) in [syms.return_stmt, syms.yield_expr] and LL[
+            0
+        ].value in ["return", "yield"]:
+            is_valid_index = is_valid_index_factory(LL)
+
+            idx = 2 if is_valid_index(1) and is_empty_par(LL[1]) else 1
+            # The next visible leaf MUST contain a string...
+            if is_valid_index(idx) and LL[idx].type == token.STRING:
+                return idx
+
+        return None
+
+    @staticmethod
+    def _else_match(LL: List[Leaf]) -> Optional[int]:
+        """
+        Returns:
+            string_idx such that @LL[string_idx] is equal to our target (i.e.
+            matched) string, if this line matches the ternary expression
+            requirements listed in the 'Requirements' section of this classes'
+            docstring.
+                OR
+            None, otherwise.
+        """
+        # If this line is apart of a ternary expression and the first leaf
+        # contains the "else" keyword...
+        if (
+            parent_type(LL[0]) == syms.test
+            and LL[0].type == token.NAME
+            and LL[0].value == "else"
+        ):
+            is_valid_index = is_valid_index_factory(LL)
+
+            idx = 2 if is_valid_index(1) and is_empty_par(LL[1]) else 1
+            # The next visible leaf MUST contain a string...
+            if is_valid_index(idx) and LL[idx].type == token.STRING:
+                return idx
+
+        return None
+
+    @staticmethod
+    def _assert_match(LL: List[Leaf]) -> Optional[int]:
+        """
+        Returns:
+            string_idx such that @LL[string_idx] is equal to our target (i.e.
+            matched) string, if this line matches the assert statement
+            requirements listed in the 'Requirements' section of this classes'
+            docstring.
+                OR
+            None, otherwise.
+        """
+        # If this line is apart of an assert statement and the first leaf
+        # contains the "assert" keyword...
+        if parent_type(LL[0]) == syms.assert_stmt and LL[0].value == "assert":
+            is_valid_index = is_valid_index_factory(LL)
+
+            for (i, leaf) in enumerate(LL):
+                # We MUST find a comma...
+                if leaf.type == token.COMMA:
+                    idx = i + 2 if is_empty_par(LL[i + 1]) else i + 1
+
+                    # That comma MUST be followed by a string...
+                    if is_valid_index(idx) and LL[idx].type == token.STRING:
+                        string_idx = idx
+
+                        # Skip the string trailer, if one exists.
+                        string_parser = StringParser()
+                        idx = string_parser.parse(LL, string_idx)
+
+                        # But no more leaves are allowed...
+                        if not is_valid_index(idx):
+                            return string_idx
+
+        return None
+
+    @staticmethod
+    def _assign_match(LL: List[Leaf]) -> Optional[int]:
+        """
+        Returns:
+            string_idx such that @LL[string_idx] is equal to our target (i.e.
+            matched) string, if this line matches the assignment statement
+            requirements listed in the 'Requirements' section of this classes'
+            docstring.
+                OR
+            None, otherwise.
+        """
+        # If this line is apart of an expression statement or is a function
+        # argument AND the first leaf contains a variable name...
+        if (
+            parent_type(LL[0]) in [syms.expr_stmt, syms.argument, syms.power]
+            and LL[0].type == token.NAME
+        ):
+            is_valid_index = is_valid_index_factory(LL)
+
+            for (i, leaf) in enumerate(LL):
+                # We MUST find either an '=' or '+=' symbol...
+                if leaf.type in [token.EQUAL, token.PLUSEQUAL]:
+                    idx = i + 2 if is_empty_par(LL[i + 1]) else i + 1
+
+                    # That symbol MUST be followed by a string...
+                    if is_valid_index(idx) and LL[idx].type == token.STRING:
+                        string_idx = idx
+
+                        # Skip the string trailer, if one exists.
+                        string_parser = StringParser()
+                        idx = string_parser.parse(LL, string_idx)
+
+                        # The next leaf MAY be a comma iff this line is apart
+                        # of a function argument...
+                        if (
+                            parent_type(LL[0]) == syms.argument
+                            and is_valid_index(idx)
+                            and LL[idx].type == token.COMMA
+                        ):
+                            idx += 1
+
+                        # But no more leaves are allowed...
+                        if not is_valid_index(idx):
+                            return string_idx
+
+        return None
+
+    @staticmethod
+    def _dict_match(LL: List[Leaf]) -> Optional[int]:
+        """
+        Returns:
+            string_idx such that @LL[string_idx] is equal to our target (i.e.
+            matched) string, if this line matches the dictionary key assignment
+            statement requirements listed in the 'Requirements' section of this
+            classes' docstring.
+                OR
+            None, otherwise.
+        """
+        # If this line is apart of a dictionary key assignment...
+        if syms.dictsetmaker in [parent_type(LL[0]), parent_type(LL[0].parent)]:
+            is_valid_index = is_valid_index_factory(LL)
+
+            for (i, leaf) in enumerate(LL):
+                # We MUST find a colon...
+                if leaf.type == token.COLON:
+                    idx = i + 2 if is_empty_par(LL[i + 1]) else i + 1
+
+                    # That colon MUST be followed by a string...
+                    if is_valid_index(idx) and LL[idx].type == token.STRING:
+                        string_idx = idx
+
+                        # Skip the string trailer, if one exists.
+                        string_parser = StringParser()
+                        idx = string_parser.parse(LL, string_idx)
+
+                        # That string MAY be followed by a comma...
+                        if is_valid_index(idx) and LL[idx].type == token.COMMA:
+                            idx += 1
+
+                        # But no more leaves are allowed...
+                        if not is_valid_index(idx):
+                            return string_idx
+
+        return None
+
+    def do_transform(self, line: Line, string_idx: int) -> Iterator[TResult[Line]]:
+        LL = line.leaves
+
+        is_valid_index = is_valid_index_factory(LL)
+        insert_str_child = insert_str_child_factory(LL[string_idx])
+
+        comma_idx = len(LL) - 1
+        ends_with_comma = False
+        if LL[comma_idx].type == token.COMMA:
+            ends_with_comma = True
+
+        leaves_to_steal_comments_from = [LL[string_idx]]
+        if ends_with_comma:
+            leaves_to_steal_comments_from.append(LL[comma_idx])
+
+        # --- First Line
+        first_line = line.clone()
+        left_leaves = LL[:string_idx]
+
+        # We have to remember to account for (possibly invisible) LPAR and RPAR
+        # leaves that already wrapped the target string. If these leaves do
+        # exist, we will replace them with our own LPAR and RPAR leaves.
+        old_parens_exist = False
+        if left_leaves and left_leaves[-1].type == token.LPAR:
+            old_parens_exist = True
+            leaves_to_steal_comments_from.append(left_leaves[-1])
+            left_leaves.pop()
+
+        append_leaves(first_line, line, left_leaves)
+
+        lpar_leaf = Leaf(token.LPAR, "(")
+        if old_parens_exist:
+            replace_child(LL[string_idx - 1], lpar_leaf)
+        else:
+            insert_str_child(lpar_leaf)
+        first_line.append(lpar_leaf)
+
+        # We throw inline comments that were originally to the right of the
+        # target string to the top line. They will now be shown to the right of
+        # the LPAR.
+        for leaf in leaves_to_steal_comments_from:
+            for comment_leaf in line.comments_after(leaf):
+                first_line.append(comment_leaf, preformatted=True)
+
+        yield Ok(first_line)
+
+        # --- Middle (String) Line
+        # We only need to yield one (possibly too long) string line, since the
+        # `StringSplitter` will break it down further if necessary.
+        string_value = LL[string_idx].value
+        string_line = Line(
+            depth=line.depth + 1,
+            inside_brackets=True,
+            should_explode=line.should_explode,
+        )
+        string_leaf = Leaf(token.STRING, string_value)
+        insert_str_child(string_leaf)
+        string_line.append(string_leaf)
+
+        old_rpar_leaf = None
+        if is_valid_index(string_idx + 1):
+            right_leaves = LL[string_idx + 1 :]
+            if ends_with_comma:
+                right_leaves.pop()
+
+            if old_parens_exist:
+                assert (
+                    right_leaves and right_leaves[-1].type == token.RPAR
+                ), "Apparently, old parentheses do NOT exist?!"
+                old_rpar_leaf = right_leaves.pop()
+
+            append_leaves(string_line, line, right_leaves)
+
+        yield Ok(string_line)
+
+        # --- Last Line
+        last_line = line.clone()
+        last_line.bracket_tracker = first_line.bracket_tracker
+
+        new_rpar_leaf = Leaf(token.RPAR, ")")
+        if old_rpar_leaf is not None:
+            replace_child(old_rpar_leaf, new_rpar_leaf)
+        else:
+            insert_str_child(new_rpar_leaf)
+        last_line.append(new_rpar_leaf)
+
+        # If the target string ended with a comma, we place this comma to the
+        # right of the RPAR on the last line.
+        if ends_with_comma:
+            comma_leaf = Leaf(token.COMMA, ",")
+            replace_child(LL[comma_idx], comma_leaf)
+            last_line.append(comma_leaf)
+
+        yield Ok(last_line)
+
+
+class StringParser:
+    """
+    A state machine that aids in parsing a string's "trailer", which can be
+    either non-existant, an old-style formatting sequence (e.g. `% varX` or `%
+    (varX, varY)`), or a method-call / attribute access (e.g. `.format(varX,
+    varY)`).
+
+    NOTE: A new StringParser object MUST be instantiated for each string
+    trailer we need to parse.
+
+    Examples:
+        We shall assume that `line` equals the `Line` object that corresponds
+        to the following line of python code:
+        ```
+        x = "Some {}.".format("String") + some_other_string
+        ```
+
+        Furthermore, we will assume that `string_idx` is some index such that:
+        ```
+        assert line.leaves[string_idx].value == "Some {}."
+        ```
+
+        The following code snippet then holds:
+        ```
+        string_parser = StringParser()
+        idx = string_parser.parse(line.leaves, string_idx)
+        assert line.leaves[idx].type == token.PLUS
+        ```
+    """
+
+    DEFAULT_TOKEN = -1
+
+    # String Parser States
+    START = 1
+    DOT = 2
+    NAME = 3
+    PERCENT = 4
+    SINGLE_FMT_ARG = 5
+    LPAR = 6
+    RPAR = 7
+    DONE = 8
+
+    # Lookup Table for Next State
+    _goto: Dict[Tuple[ParserState, NodeType], ParserState] = {
+        # A string trailer may start with '.' OR '%'.
+        (START, token.DOT): DOT,
+        (START, token.PERCENT): PERCENT,
+        (START, DEFAULT_TOKEN): DONE,
+        # A '.' MUST be followed by an attribute or method name.
+        (DOT, token.NAME): NAME,
+        # A method name MUST be followed by an '(', whereas an attribute name
+        # is the last symbol in the string trailer.
+        (NAME, token.LPAR): LPAR,
+        (NAME, DEFAULT_TOKEN): DONE,
+        # A '%' symbol can be followed by an '(' or a single argument (e.g. a
+        # string or variable name).
+        (PERCENT, token.LPAR): LPAR,
+        (PERCENT, DEFAULT_TOKEN): SINGLE_FMT_ARG,
+        # If a '%' symbol is followed by a single argument, that argument is
+        # the last leaf in the string trailer.
+        (SINGLE_FMT_ARG, DEFAULT_TOKEN): DONE,
+        # If present, a ')' symbol is the last symbol in a string trailer.
+        # (NOTE: LPARS and nested RPARS are not included in this lookup table,
+        # since they are treated as a special case by the parsing logic in this
+        # classes' implementation.)
+        (RPAR, DEFAULT_TOKEN): DONE,
+    }
+
+    def __init__(self) -> None:
+        self._state = self.START
+        self._unmatched_lpars = 0
+
+    def parse(self, leaves: List[Leaf], string_idx: int) -> int:
+        """
+        Pre-conditions:
+            * @leaves[@string_idx].type == token.STRING
+
+        Returns:
+            The index directly after the last leaf which is apart of the string
+            trailer, if a "trailer" exists.
+                OR
+            @string_idx + 1, if no string "trailer" exists.
+        """
+        assert leaves[string_idx].type == token.STRING
+
+        idx = string_idx + 1
+        while idx < len(leaves) and self._next_state(leaves[idx]):
+            idx += 1
+        return idx
+
+    def _next_state(self, leaf: Leaf) -> bool:
+        """
+        Pre-conditions:
+            * On the first call to this function, @leaf MUST be the leaf that
+            was directly after the string leaf in question (e.g. if our target
+            string is `line.leaves[i]` then the first call to this method must
+            be `line.leaves[i + 1]`).
+            * On the next call to this function, the leaf paramater passed in
+            MUST be the leaf directly following @leaf.
+
+        Returns:
+            True iff @leaf is apart of the string's trailer.
+        """
+        # We ignore empty LPAR or RPAR leaves.
+        if is_empty_par(leaf):
+            return True
+
+        next_token = leaf.type
+        if next_token == token.LPAR:
+            self._unmatched_lpars += 1
+
+        current_state = self._state
+
+        # The LPAR parser state is a special case. We will return True until we
+        # find the matching RPAR token.
+        if current_state == self.LPAR:
+            if next_token == token.RPAR:
+                self._unmatched_lpars -= 1
+                if self._unmatched_lpars == 0:
+                    self._state = self.RPAR
+        # Otherwise, we use a lookup table to determine the next state.
+        else:
+            # If the lookup table matches the current state to the next
+            # token, we use the lookup table.
+            if (current_state, next_token) in self._goto:
+                self._state = self._goto[current_state, next_token]
+            else:
+                # Otherwise, we check if a the current state was assigned a
+                # default.
+                if (current_state, self.DEFAULT_TOKEN) in self._goto:
+                    self._state = self._goto[current_state, self.DEFAULT_TOKEN]
+                # If no default has been assigned, then this parser has a logic
+                # error.
+                else:
+                    raise RuntimeError(f"{self.__class__.__name__} LOGIC ERROR!")
+
+            if self._state == self.DONE:
+                return False
+
+        return True
+
+
+def TErr(err_msg: str) -> Err[CannotTransform]:
+    """(T)ransform Err
+
+    Convenience function used when working with the TResult type.
+    """
+    cant_transform = CannotTransform(err_msg)
+    return Err(cant_transform)
+
+
+def contains_pragma_comment(comment_list: List[Leaf]) -> bool:
+    """
+    Returns:
+        True iff one of the comments in @comment_list is a pragma used by one
+        of the more common static analysis tools for python (e.g. mypy, flake8,
+        pylint).
+    """
+    for comment in comment_list:
+        if comment.value.startswith(("# type:", "# noqa", "# pylint:")):
+            return True
+
+    return False
+
+
+def insert_str_child_factory(string_leaf: Leaf) -> Callable[[LN], None]:
+    """
+    Factory for a convenience function that is used to orphan @string_leaf
+    and then insert multiple new leaves into the same part of the node
+    structure that @string_leaf had originally occupied.
+
+    Examples:
+        Let `string_leaf = Leaf(token.STRING, '"foo"')` and `N =
+        string_leaf.parent`. Assume the node `N` has the following
+        original structure:
+
+        Node(
+            expr_stmt, [
+                Leaf(NAME, 'x'),
+                Leaf(EQUAL, '='),
+                Leaf(STRING, '"foo"'),
+            ]
+        )
+
+        We then run the code snippet shown below.
+        ```
+        insert_str_child = insert_str_child_factory(string_leaf)
+
+        lpar = Leaf(token.LPAR, '(')
+        insert_str_child(lpar)
+
+        bar = Leaf(token.STRING, '"bar"')
+        insert_str_child(bar)
+
+        rpar = Leaf(token.RPAR, ')')
+        insert_str_child(rpar)
+        ```
+
+        After which point, it follows that `string_leaf.parent is None` and
+        the node `N` now has the following structure:
+
+        Node(
+            expr_stmt, [
+                Leaf(NAME, 'x'),
+                Leaf(EQUAL, '='),
+                Leaf(LPAR, '('),
+                Leaf(STRING, '"bar"'),
+                Leaf(RPAR, ')'),
+            ]
+        )
+    """
+    string_parent = string_leaf.parent
+    string_child_idx = string_leaf.remove()
+
+    def insert_str_child(child: LN) -> None:
+        nonlocal string_child_idx
+
+        assert string_parent is not None
+        assert string_child_idx is not None
+
+        string_parent.insert_child(string_child_idx, child)
+        string_child_idx += 1
+
+    return insert_str_child
+
+
+def has_triple_quotes(string: str) -> bool:
+    """
+    Returns:
+        True iff @string starts with three quotation characters.
+    """
+    raw_string = string.lstrip(STRING_PREFIX_CHARS)
+    return raw_string[:3] in {'"""', "'''"}
+
+
+def parent_type(node: Optional[LN]) -> Optional[NodeType]:
+    """
+    Returns:
+        @node.parent.type, if @node is not None and has a parent.
+            OR
+        None, otherwise.
+    """
+    if node is None or node.parent is None:
+        return None
+
+    return node.parent.type
+
+
+def is_empty_par(leaf: Leaf) -> bool:
+    return is_empty_lpar(leaf) or is_empty_rpar(leaf)
+
+
+def is_empty_lpar(leaf: Leaf) -> bool:
+    return leaf.type == token.LPAR and leaf.value == ""
+
+
+def is_empty_rpar(leaf: Leaf) -> bool:
+    return leaf.type == token.RPAR and leaf.value == ""
+
+
+def is_valid_index_factory(seq: Sequence[Any]) -> Callable[[int], bool]:
+    """
+    Examples:
+        ```
+        my_list = [1, 2, 3]
+
+        is_valid_index = is_valid_index_factory(my_list)
+
+        assert is_valid_index(0)
+        assert is_valid_index(2)
+
+        assert not is_valid_index(3)
+        assert not is_valid_index(-1)
+        ```
+    """
+
+    def is_valid_index(idx: int) -> bool:
+        """
+        Returns:
+            True iff @idx is positive AND seq[@idx] does NOT raise an
+            IndexError.
+        """
+        return 0 <= idx < len(seq)
+
+    return is_valid_index
+
+
+def line_to_string(line: Line) -> str:
+    """Returns the string representation of @line.
+
+    WARNING: This is known to be computationally expensive.
+    """
+    return str(line).strip("\n")
+
+
+def append_leaves(new_line: Line, old_line: Line, leaves: List[Leaf]) -> None:
+    """
+    Append leaves (taken from @old_line) to @new_line, making sure to fix the
+    underlying Node structure where appropriate.
+
+    All of the leaves in @leaves are duplicated. The duplicates are then
+    appended to @new_line and used to replace their originals in the underlying
+    Node structure. Any comments attatched to the old leaves are reattached to
+    the new leaves.
+
+    Pre-conditions:
+        set(@leaves) is a subset of set(@old_line.leaves).
+    """
+    for old_leaf in leaves:
+        assert old_leaf in old_line.leaves
+
+        new_leaf = Leaf(old_leaf.type, old_leaf.value)
+        replace_child(old_leaf, new_leaf)
+        new_line.append(new_leaf)
+
+        for comment_leaf in old_line.comments_after(old_leaf):
+            new_line.append(comment_leaf, preformatted=True)
+
+
+def replace_child(old_child: LN, new_child: LN) -> None:
+    """
+    Side Effects:
+        * If @old_child.parent is set, replace @old_child with @new_child in
+        @old_child's underlying Node structure.
+            OR
+        * Otherwise, this function does nothing.
+    """
+    parent = old_child.parent
+    if not parent:
+        return
+
+    child_idx = old_child.remove()
+    if child_idx is not None:
+        parent.insert_child(child_idx, new_child)
+
+
+def get_string_prefix(string: str) -> str:
+    """
+    Pre-conditions:
+        * assert_is_leaf_string(@string)
+
+    Returns:
+        @string's prefix (e.g. '', 'r', 'f', or 'rf').
+    """
+    assert_is_leaf_string(string)
+
+    prefix = ""
+    prefix_idx = 0
+    while string[prefix_idx] in STRING_PREFIX_CHARS:
+        prefix += string[prefix_idx].lower()
+        prefix_idx += 1
+
+    return prefix
+
+
+def assert_is_leaf_string(string: str) -> None:
+    """
+    Checks the pre-condition that @string has the format that you would expect
+    of `leaf.value` where `leaf` is some Leaf such that `leaf.type ==
+    token.STRING`. A more precise description of the pre-conditions that are
+    checked are listed below.
+
+    Pre-conditions:
+        * @string starts with either ', ", <prefix>', or <prefix>" where
+        `set(<prefix>)` is some subset of `set(STRING_PREFIX_CHARS)`.
+        * @string ends with a quote character (' or ").
+
+    Raises:
+        AssertionError(...) if the pre-conditions listed above are not
+        satisfied.
+    """
+    dquote_idx = string.find('"')
+    squote_idx = string.find("'")
+    if -1 in [dquote_idx, squote_idx]:
+        quote_idx = max(dquote_idx, squote_idx)
+    else:
+        quote_idx = min(squote_idx, dquote_idx)
+
+    assert (
+        0 <= quote_idx < len(string) - 1
+    ), f"{string!r} is missing a starting quote character (' or \")."
+    assert string[-1] in (
+        "'",
+        '"',
+    ), f"{string!r} is missing an ending quote character (' or \")."
+    assert set(string[:quote_idx]).issubset(
+        set(STRING_PREFIX_CHARS)
+    ), f"{set(string[:quote_idx])} is NOT a subset of {set(STRING_PREFIX_CHARS)}."
+
+
+def left_hand_split(line: Line, _features: Collection[Feature] = ()) -> Iterator[Line]:
+    """Split line into many lines, starting with the first matching bracket pair.
+
+    Note: this usually looks weird, only use this for function definitions.
+    Prefer RHS otherwise.  This is why this function is not symmetrical with
+    :func:`right_hand_split` which also handles optional parentheses.
+    """
+    tail_leaves: List[Leaf] = []
+    body_leaves: List[Leaf] = []
+    head_leaves: List[Leaf] = []
+    current_leaves = head_leaves
+    matching_bracket: Optional[Leaf] = None
+    for leaf in line.leaves:
+        if (
+            current_leaves is body_leaves
+            and leaf.type in CLOSING_BRACKETS
+            and leaf.opening_bracket is matching_bracket
+        ):
+            current_leaves = tail_leaves if body_leaves else head_leaves
+        current_leaves.append(leaf)
+        if current_leaves is head_leaves:
+            if leaf.type in OPENING_BRACKETS:
+                matching_bracket = leaf
+                current_leaves = body_leaves
+    if not matching_bracket:
+        raise CannotSplit("No brackets found")
+
+    head = bracket_split_build_line(head_leaves, line, matching_bracket)
+    body = bracket_split_build_line(body_leaves, line, matching_bracket, is_body=True)
+    tail = bracket_split_build_line(tail_leaves, line, matching_bracket)
+    bracket_split_succeeded_or_raise(head, body, tail)
+    for result in (head, body, tail):
+        if result:
+            yield result
+
+
+def right_hand_split(
+    line: Line,
+    line_length: int,
+    features: Collection[Feature] = (),
+    omit: Collection[LeafID] = (),
+) -> Iterator[Line]:
+    """Split line into many lines, starting with the last matching bracket pair.
+
+    If the split was by optional parentheses, attempt splitting without them, too.
+    `omit` is a collection of closing bracket IDs that shouldn't be considered for
+    this split.
+
+    Note: running this function modifies `bracket_depth` on the leaves of `line`.
+    """
+    tail_leaves: List[Leaf] = []
+    body_leaves: List[Leaf] = []
+    head_leaves: List[Leaf] = []
+    current_leaves = tail_leaves
+    opening_bracket: Optional[Leaf] = None
+    closing_bracket: Optional[Leaf] = None
+    for leaf in reversed(line.leaves):
+        if current_leaves is body_leaves:
+            if leaf is opening_bracket:
+                current_leaves = head_leaves if body_leaves else tail_leaves
+        current_leaves.append(leaf)
+        if current_leaves is tail_leaves:
+            if leaf.type in CLOSING_BRACKETS and id(leaf) not in omit:
+                opening_bracket = leaf.opening_bracket
                 closing_bracket = leaf
                 current_leaves = body_leaves
     if not (opening_bracket and closing_bracket and head_leaves):
@@ -2565,10 +4777,10 @@ def right_hand_split(
 
             elif head.contains_multiline_strings() or tail.contains_multiline_strings():
                 raise CannotSplit(
-                    "The current optional pair of parentheses is bound to fail to "
-                    "satisfy the splitting algorithm because the head or the tail "
-                    "contains multiline strings which by definition never fit one "
-                    "line."
+                    "The current optional pair of parentheses is bound to fail to"
+                    " satisfy the splitting algorithm because the head or the tail"
+                    " contains multiline strings which by definition never fit one"
+                    " line."
                 )
 
     ensure_visible(opening_bracket)
@@ -2599,8 +4811,8 @@ def bracket_split_succeeded_or_raise(head: Line, body: Line, tail: Line) -> None
 
         elif tail_len < 3:
             raise CannotSplit(
-                f"Splitting brackets on an empty body to save "
-                f"{tail_len} characters is not worth it"
+                f"Splitting brackets on an empty body to save {tail_len} characters is"
+                " not worth it"
             )
 
 
@@ -2620,20 +4832,22 @@ def bracket_split_build_line(
             # Since body is a new indent level, remove spurious leading whitespace.
             normalize_prefix(leaves[0], inside_brackets=True)
             # Ensure a trailing comma for imports and standalone function arguments, but
-            # be careful not to add one after any comments.
-            no_commas = original.is_def and not any(
-                l.type == token.COMMA for l in leaves
+            # be careful not to add one after any comments or within type annotations.
+            no_commas = (
+                original.is_def
+                and opening_bracket.value == "("
+                and not any(l.type == token.COMMA for l in leaves)
             )
 
             if original.is_import or no_commas:
                 for i in range(len(leaves) - 1, -1, -1):
                     if leaves[i].type == STANDALONE_COMMENT:
                         continue
-                    elif leaves[i].type == token.COMMA:
-                        break
-                    else:
+
+                    if leaves[i].type != token.COMMA:
                         leaves.insert(i + 1, Leaf(token.COMMA, ","))
-                        break
+                    break
+
     # Populate the line
     for leaf in leaves:
         result.append(leaf, preformatted=True)
@@ -2644,7 +4858,7 @@ def bracket_split_build_line(
     return result
 
 
-def dont_increase_indentation(split_func: SplitFunc) -> SplitFunc:
+def dont_increase_indentation(split_func: Transformer) -> Transformer:
     """Normalize prefix of the first leaf in every line returned by `split_func`.
 
     This is a decorator over relevant split functions.
@@ -2779,9 +4993,7 @@ def is_type_comment(leaf: Leaf, suffix: str = "") -> bool:
     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:" + suffix
-    )
+    return t in {token.COMMENT, STANDALONE_COMMENT} and v.startswith("# type:" + suffix)
 
 
 def normalize_prefix(leaf: Leaf, *, inside_brackets: bool) -> None:
@@ -2809,10 +5021,10 @@ def normalize_string_prefix(leaf: Leaf, remove_u_prefix: bool = False) -> None:
 
     Note: Mutates its argument.
     """
-    match = re.match(r"^([furbFURB]*)(.*)$", leaf.value, re.DOTALL)
+    match = re.match(r"^([" + STRING_PREFIX_CHARS + r"]*)(.*)$", leaf.value, re.DOTALL)
     assert match is not None, f"failed to match string {leaf.value!r}"
     orig_prefix = match.group(1)
-    new_prefix = orig_prefix.lower()
+    new_prefix = orig_prefix.replace("F", "f").replace("B", "b").replace("U", "u")
     if remove_u_prefix:
         new_prefix = new_prefix.replace("u", "")
     leaf.value = f"{new_prefix}{match.group(2)}"
@@ -2826,7 +5038,7 @@ def normalize_string_quotes(leaf: Leaf) -> None:
 
     Note: Mutates its argument.
     """
-    value = leaf.value.lstrip("furbFURB")
+    value = leaf.value.lstrip(STRING_PREFIX_CHARS)
     if value[:3] == '"""':
         return
 
@@ -2879,6 +5091,7 @@ def normalize_string_quotes(leaf: Leaf) -> None:
             if "\\" in str(m):
                 # Do not introduce backslashes in interpolated expressions
                 return
+
     if new_quote == '"""' and new_body[-1:] == '"':
         # edge case:
         new_body = new_body[:-1] + '\\"'
@@ -2951,9 +5164,13 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
         if pc.value in FMT_OFF:
             # This `node` has a prefix with `# fmt: off`, don't mess with parens.
             return
-
     check_lpar = False
     for index, child in enumerate(list(node.children)):
+        # Fixes a bug where invisible parens are not properly stripped from
+        # assignment statements that contain type annotations.
+        if isinstance(child, Node) and child.type == syms.annassign:
+            normalize_invisible_parens(child, parens_after=parens_after)
+
         # Add parentheses around long tuple unpacking in assignments.
         if (
             index == 0
@@ -2965,26 +5182,12 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
         if check_lpar:
             if is_walrus_assignment(child):
                 continue
-            if child.type == syms.atom:
-                # 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
-                    node.insert_child(index, Node(syms.atom, [lpar, child, rpar]))
+            if child.type == syms.atom:
+                if maybe_make_parens_invisible_in_atom(child, parent=node):
+                    wrap_in_parentheses(node, child, visible=False)
             elif is_one_tuple(child):
-                # wrap child in visible parentheses
-                lpar = Leaf(token.LPAR, "(")
-                rpar = Leaf(token.RPAR, ")")
-                child.remove()
-                node.insert_child(index, Node(syms.atom, [lpar, child, rpar]))
+                wrap_in_parentheses(node, child, visible=True)
             elif node.type == syms.import_from:
                 # "import from" nodes store parentheses directly as part of
                 # the statement
@@ -2999,15 +5202,7 @@ def normalize_invisible_parens(node: Node, parens_after: Set[str]) -> None:
                 break
 
             elif not (isinstance(child, Leaf) and is_multiline_string(child)):
-                # wrap child in invisible parentheses
-                lpar = Leaf(token.LPAR, "")
-                rpar = Leaf(token.RPAR, "")
-                index = child.remove() or 0
-                prefix = child.prefix
-                child.prefix = ""
-                new_child = Node(syms.atom, [lpar, child, rpar])
-                new_child.prefix = prefix
-                node.insert_child(index, new_child)
+                wrap_in_parentheses(node, child, visible=False)
 
         check_lpar = isinstance(child, Leaf) and child.value in parens_after
 
@@ -3051,7 +5246,7 @@ def convert_one_fmt_off_pair(node: Node) -> bool:
                     # That happens when one of the `ignored_nodes` ended with a NEWLINE
                     # leaf (possibly followed by a DEDENT).
                     hidden_value = hidden_value[:-1]
-                first_idx = None
+                first_idx: Optional[int] = None
                 for ignored in ignored_nodes:
                     index = ignored.remove()
                     if first_idx is None:
@@ -3080,13 +5275,54 @@ def generate_ignored_nodes(leaf: Leaf) -> Iterator[LN]:
     """
     container: Optional[LN] = container_of(leaf)
     while container is not None and container.type != token.ENDMARKER:
-        for comment in list_comments(container.prefix, is_endmarker=False):
-            if comment.value in FMT_ON:
-                return
+        if is_fmt_on(container):
+            return
+
+        # fix for fmt: on in children
+        if contains_fmt_on_at_column(container, leaf.column):
+            for child in container.children:
+                if contains_fmt_on_at_column(child, leaf.column):
+                    return
+                yield child
+        else:
+            yield container
+            container = container.next_sibling
+
+
+def is_fmt_on(container: LN) -> bool:
+    """Determine whether formatting is switched on within a container.
+    Determined by whether the last `# fmt:` comment is `on` or `off`.
+    """
+    fmt_on = False
+    for comment in list_comments(container.prefix, is_endmarker=False):
+        if comment.value in FMT_ON:
+            fmt_on = True
+        elif comment.value in FMT_OFF:
+            fmt_on = False
+    return fmt_on
+
+
+def contains_fmt_on_at_column(container: LN, column: int) -> bool:
+    """Determine if children at a given column have formatting switched on."""
+    for child in container.children:
+        if (
+            isinstance(child, Node)
+            and first_leaf_column(child) == column
+            or isinstance(child, Leaf)
+            and child.column == column
+        ):
+            if is_fmt_on(child):
+                return True
+
+    return False
 
-        yield container
 
-        container = container.next_sibling
+def first_leaf_column(node: Node) -> Optional[int]:
+    """Returns the column of the first leaf child of a node."""
+    for child in node.children:
+        if isinstance(child, Leaf):
+            return child.column
+    return None
 
 
 def maybe_make_parens_invisible_in_atom(node: LN, parent: LN) -> bool:
@@ -3159,6 +5395,7 @@ def unwrap_singleton_parenthesis(node: LN) -> Optional[LN]:
     Parenthesis can be optional. Returns None otherwise"""
     if len(node.children) != 3:
         return None
+
     lpar, wrapped, rpar = node.children
     if not (lpar.type == token.LPAR and rpar.type == token.RPAR):
         return None
@@ -3166,6 +5403,24 @@ def unwrap_singleton_parenthesis(node: LN) -> Optional[LN]:
     return wrapped
 
 
+def wrap_in_parentheses(parent: Node, child: LN, *, visible: bool = True) -> None:
+    """Wrap `child` in parentheses.
+
+    This replaces `child` with an atom holding the parentheses and the old
+    child.  That requires moving the prefix.
+
+    If `visible` is False, the leaves will be valueless (and thus invisible).
+    """
+    lpar = Leaf(token.LPAR, "(" if visible else "")
+    rpar = Leaf(token.RPAR, ")" if visible else "")
+    prefix = child.prefix
+    child.prefix = ""
+    index = child.remove() or 0
+    new_child = Node(syms.atom, [lpar, child, rpar])
+    new_child.prefix = prefix
+    parent.insert_child(index, new_child)
+
+
 def is_one_tuple(node: LN) -> bool:
     """Return True if `node` holds a tuple with one element, with or without parens."""
     if node.type == syms.atom:
@@ -3234,8 +5489,7 @@ def is_vararg(leaf: Leaf, within: Set[NodeType]) -> bool:
 
 def is_multiline_string(leaf: Leaf) -> bool:
     """Return True if `leaf` is a multiline string that actually spans many lines."""
-    value = leaf.value.lstrip("furbFURB")
-    return value[:3] in {'"""', "'''"} and "\n" in value
+    return has_triple_quotes(leaf.value) and "\n" in leaf.value
 
 
 def is_stub_suite(node: Node) -> bool:
@@ -3398,8 +5652,8 @@ def generate_trailers_to_omit(line: Line, line_length: int) -> Iterator[Set[Leaf
     yield omit
 
     length = 4 * line.depth
-    opening_bracket = None
-    closing_bracket = None
+    opening_bracket: Optional[Leaf] = None
+    closing_bracket: Optional[Leaf] = None
     inner_brackets: Set[LeafID] = set()
     for index, leaf, leaf_length in enumerate_with_length(line, reversed=True):
         length += leaf_length
@@ -3443,19 +5697,23 @@ def get_future_imports(node: Node) -> Set[str]:
             if isinstance(child, Leaf):
                 if child.type == token.NAME:
                     yield child.value
+
             elif child.type == syms.import_as_name:
                 orig_name = child.children[0]
                 assert isinstance(orig_name, Leaf), "Invalid syntax parsing imports"
                 assert orig_name.type == token.NAME, "Invalid syntax parsing imports"
                 yield orig_name.value
+
             elif child.type == syms.import_as_names:
                 yield from get_imports_from_children(child.children)
+
             else:
                 raise AssertionError("Invalid syntax parsing imports")
 
     for child in node.children:
         if child.type != syms.simple_stmt:
             break
+
         first_child = child.children[0]
         if isinstance(first_child, Leaf):
             # Continue looking if we see a docstring; otherwise stop.
@@ -3465,24 +5723,39 @@ def get_future_imports(node: Node) -> Set[str]:
                 and child.children[1].type == token.NEWLINE
             ):
                 continue
-            else:
-                break
+
+            break
+
         elif first_child.type == syms.import_from:
             module_name = first_child.children[1]
             if not isinstance(module_name, Leaf) or module_name.value != "__future__":
                 break
+
             imports |= set(get_imports_from_children(first_child.children[3:]))
         else:
             break
+
     return imports
 
 
-def gen_python_files_in_dir(
-    path: Path,
+@lru_cache()
+def get_gitignore(root: Path) -> PathSpec:
+    """ Return a PathSpec matching gitignore content if present."""
+    gitignore = root / ".gitignore"
+    lines: List[str] = []
+    if gitignore.is_file():
+        with gitignore.open() as gf:
+            lines = gf.readlines()
+    return PathSpec.from_lines("gitwildmatch", lines)
+
+
+def gen_python_files(
+    paths: Iterable[Path],
     root: Path,
-    include: Pattern[str],
-    exclude: Pattern[str],
+    include: Optional[Pattern[str]],
+    exclude_regexes: Iterable[Pattern[str]],
     report: "Report",
+    gitignore: PathSpec,
 ) -> Iterator[Path]:
     """Generate all files under `path` whose paths are not excluded by the
     `exclude` regex, but are included by the `include` regex.
@@ -3492,9 +5765,13 @@ def gen_python_files_in_dir(
     `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 paths:
+        # Then ignore with `exclude` option.
         try:
-            normalized_path = "/" + child.resolve().relative_to(root).as_posix()
+            normalized_path = child.resolve().relative_to(root).as_posix()
+        except OSError as e:
+            report.path_ignored(child, f"cannot be read because {e}")
+            continue
         except ValueError:
             if child.is_symlink():
                 report.path_ignored(
@@ -3504,18 +5781,32 @@ def gen_python_files_in_dir(
 
             raise
 
+        # First ignore files matching .gitignore
+        if gitignore.match_file(normalized_path):
+            report.path_ignored(child, "matches the .gitignore file content")
+            continue
+
+        normalized_path = "/" + normalized_path
         if child.is_dir():
             normalized_path += "/"
-        exclude_match = exclude.search(normalized_path)
-        if exclude_match and exclude_match.group(0):
-            report.path_ignored(child, f"matches the --exclude regular expression")
+
+        is_excluded = False
+        for exclude in exclude_regexes:
+            exclude_match = exclude.search(normalized_path) if exclude else None
+            if exclude_match and exclude_match.group(0):
+                report.path_ignored(child, "matches the --exclude regular expression")
+                is_excluded = True
+                break
+        if is_excluded:
             continue
 
         if child.is_dir():
-            yield from gen_python_files_in_dir(child, root, include, exclude, report)
+            yield from gen_python_files(
+                child.iterdir(), root, include, exclude_regexes, report, gitignore
+            )
 
         elif child.is_file():
-            include_match = include.search(normalized_path)
+            include_match = include.search(normalized_path) if include else True
             if include_match:
                 yield child
 
@@ -3538,7 +5829,7 @@ def find_project_root(srcs: Iterable[str]) -> Path:
         # 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():
+        if (directory / ".git").exists():
             return directory
 
         if (directory / ".hg").is_dir():
@@ -3555,6 +5846,7 @@ class Report:
     """Provides a reformatting counter. Can be rendered with `str(report)`."""
 
     check: bool = False
+    diff: bool = False
     quiet: bool = False
     verbose: bool = False
     change_count: int = 0
@@ -3564,7 +5856,7 @@ class Report:
     def done(self, src: Path, changed: Changed) -> None:
         """Increment the counter for successful reformatting. Write out a message."""
         if changed is Changed.YES:
-            reformatted = "would reformat" if self.check else "reformatted"
+            reformatted = "would reformat" if self.check or self.diff else "reformatted"
             if self.verbose or not self.quiet:
                 out(f"{reformatted} {src}")
             self.change_count += 1
@@ -3610,7 +5902,7 @@ class Report:
 
         Use `click.unstyle` to remove colors.
         """
-        if self.check:
+        if self.check or self.diff:
             reformatted = "would be reformatted"
             unchanged = "would be left unchanged"
             failed = "would fail to reformat"
@@ -3658,70 +5950,84 @@ def _fixup_ast_constants(
     node: Union[ast.AST, ast3.AST, ast27.AST]
 ) -> Union[ast.AST, ast3.AST, ast27.AST]:
     """Map ast nodes deprecated in 3.8 to Constant."""
-    # casts are required until this is released:
-    # https://github.com/python/typeshed/pull/3142
     if isinstance(node, (ast.Str, ast3.Str, ast27.Str, ast.Bytes, ast3.Bytes)):
-        return cast(ast.AST, ast.Constant(value=node.s))
-    elif isinstance(node, (ast.Num, ast3.Num, ast27.Num)):
-        return cast(ast.AST, ast.Constant(value=node.n))
-    elif isinstance(node, (ast.NameConstant, ast3.NameConstant)):
-        return cast(ast.AST, ast.Constant(value=node.value))
+        return ast.Constant(value=node.s)
+
+    if isinstance(node, (ast.Num, ast3.Num, ast27.Num)):
+        return ast.Constant(value=node.n)
+
+    if isinstance(node, (ast.NameConstant, ast3.NameConstant)):
+        return ast.Constant(value=node.value)
+
     return node
 
 
-def assert_equivalent(src: str, dst: str) -> None:
-    """Raise AssertionError if `src` and `dst` aren't equivalent."""
+def _stringify_ast(
+    node: Union[ast.AST, ast3.AST, ast27.AST], depth: int = 0
+) -> Iterator[str]:
+    """Simple visitor generating strings to compare ASTs by content."""
 
-    def _v(node: Union[ast.AST, ast3.AST, ast27.AST], depth: int = 0) -> Iterator[str]:
-        """Simple visitor generating strings to compare ASTs by content."""
+    node = _fixup_ast_constants(node)
 
-        node = _fixup_ast_constants(node)
+    yield f"{'  ' * depth}{node.__class__.__name__}("
 
-        yield f"{'  ' * depth}{node.__class__.__name__}("
+    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
 
-        for field in sorted(node._fields):
-            # 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
+        try:
+            value = getattr(node, field)
+        except AttributeError:
+            continue
 
-            try:
-                value = getattr(node, field)
-            except AttributeError:
-                continue
+        yield f"{'  ' * (depth+1)}{field}="
 
-            yield f"{'  ' * (depth+1)}{field}="
+        if isinstance(value, list):
+            for item in value:
+                # Ignore nested tuples within del statements, because we may insert
+                # 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))
+                ):
+                    for item in item.elts:
+                        yield from _stringify_ast(item, depth + 2)
 
-            if isinstance(value, list):
-                for item in value:
-                    # Ignore nested tuples within del statements, because we may insert
-                    # 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))
-                    ):
-                        for item in item.elts:
-                            yield from _v(item, depth + 2)
-                    elif isinstance(item, (ast.AST, ast3.AST, ast27.AST)):
-                        yield from _v(item, depth + 2)
+                elif isinstance(item, (ast.AST, ast3.AST, ast27.AST)):
+                    yield from _stringify_ast(item, depth + 2)
 
-            elif isinstance(value, (ast.AST, ast3.AST, ast27.AST)):
-                yield from _v(value, depth + 2)
+        elif isinstance(value, (ast.AST, ast3.AST, ast27.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
+            if (
+                isinstance(node, ast.Constant)
+                and field == "value"
+                and isinstance(value, str)
+            ):
+                normalized = re.sub(r"\n[ \t]+", "\n ", value)
             else:
-                yield f"{'  ' * (depth+2)}{value!r},  # {value.__class__.__name__}"
+                normalized = value
+            yield f"{'  ' * (depth+2)}{normalized!r},  # {value.__class__.__name__}"
 
-        yield f"{'  ' * depth})  # /{node.__class__.__name__}"
+    yield f"{'  ' * depth})  # /{node.__class__.__name__}"
 
+
+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.  "
-            f"AST error message: {exc}"
+            "cannot use --safe with this file; failed to parse source file.  AST"
+            f" error message: {exc}"
         )
 
     try:
@@ -3729,24 +6035,23 @@ def assert_equivalent(src: str, dst: str) -> 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: {exc}. "
-            f"Please report a bug on https://github.com/psf/black/issues.  "
-            f"This invalid output might be helpful: {log}"
+            f"INTERNAL ERROR: Black produced invalid code: {exc}. Please report a bug"
+            " on https://github.com/psf/black/issues.  This invalid output might be"
+            f" helpful: {log}"
         ) from None
 
-    src_ast_str = "\n".join(_v(src_ast))
-    dst_ast_str = "\n".join(_v(dst_ast))
+    src_ast_str = "\n".join(_stringify_ast(src_ast))
+    dst_ast_str = "\n".join(_stringify_ast(dst_ast))
     if src_ast_str != dst_ast_str:
         log = dump_to_file(diff(src_ast_str, dst_ast_str, "src", "dst"))
         raise AssertionError(
-            f"INTERNAL ERROR: Black produced code that is not equivalent to "
-            f"the source.  "
-            f"Please report a bug on https://github.com/psf/black/issues.  "
-            f"This diff might be helpful: {log}"
+            "INTERNAL ERROR: Black produced code that is not equivalent to the"
+            " source.  Please report a bug on https://github.com/psf/black/issues. "
+            f" This diff might be helpful: {log}"
         ) from None
 
 
-def assert_stable(src: str, dst: str, mode: FileMode) -> 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)
     if dst != newdst:
@@ -3755,13 +6060,13 @@ def assert_stable(src: str, dst: str, mode: FileMode) -> None:
             diff(dst, newdst, "first pass", "second pass"),
         )
         raise AssertionError(
-            f"INTERNAL ERROR: Black produced different code on the second pass "
-            f"of the formatter.  "
-            f"Please report a bug on https://github.com/psf/black/issues.  "
-            f"This diff might be helpful: {log}"
+            "INTERNAL ERROR: Black produced different code on the second pass of the"
+            " formatter.  Please report a bug on https://github.com/psf/black/issues."
+            f"  This diff might be helpful: {log}"
         ) from None
 
 
+@mypyc_attr(patchable=True)
 def dump_to_file(*output: str) -> str:
     """Dump `output` to a temporary file. Return path to the file."""
     with tempfile.NamedTemporaryFile(
@@ -3776,8 +6081,10 @@ def dump_to_file(*output: str) -> str:
 
 @contextmanager
 def nullcontext() -> Iterator[None]:
-    """Return context manager that does nothing.
-    Similar to `nullcontext` from python 3.7"""
+    """Return an empty context manager.
+
+    To be used like `nullcontext` in Python 3.7.
+    """
     yield
 
 
@@ -3785,14 +6092,14 @@ def diff(a: str, b: str, a_name: str, b_name: str) -> str:
     """Return a unified diff string between strings `a` and `b`."""
     import difflib
 
-    a_lines = [line + "\n" for line in a.split("\n")]
-    b_lines = [line + "\n" for line in b.split("\n")]
+    a_lines = [line + "\n" for line in a.splitlines()]
+    b_lines = [line + "\n" for line in b.splitlines()]
     return "".join(
         difflib.unified_diff(a_lines, b_lines, fromfile=a_name, tofile=b_name, n=5)
     )
 
 
-def cancel(tasks: Iterable[asyncio.Task]) -> None:
+def cancel(tasks: Iterable["asyncio.Task[Any]"]) -> None:
     """asyncio signal handler that cancels all `tasks` and reports to stderr."""
     err("Aborted!")
     for task in tasks:
@@ -3881,7 +6188,7 @@ def is_line_short_enough(line: Line, *, line_length: int, line_str: str = "") ->
     Uses the provided `line_str` rendering, if any, otherwise computes a new one.
     """
     if not line_str:
-        line_str = str(line).strip("\n")
+        line_str = line_to_string(line)
     return (
         len(line_str) <= line_length
         and "\n" not in line_str  # multiline strings
@@ -4014,11 +6321,11 @@ def can_omit_invisible_parens(line: Line, line_length: int) -> bool:
     return False
 
 
-def get_cache_file(mode: FileMode) -> Path:
+def get_cache_file(mode: Mode) -> Path:
     return CACHE_DIR / f"cache.{mode.get_cache_key()}.pickle"
 
 
-def read_cache(mode: FileMode) -> Cache:
+def read_cache(mode: Mode) -> 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.
@@ -4030,7 +6337,7 @@ def read_cache(mode: FileMode) -> Cache:
     with cache_file.open("rb") as fobj:
         try:
             cache: Cache = pickle.load(fobj)
-        except pickle.UnpicklingError:
+        except (pickle.UnpicklingError, ValueError):
             return {}
 
     return cache
@@ -4058,14 +6365,14 @@ def filter_cached(cache: Cache, sources: Iterable[Path]) -> Tuple[Set[Path], Set
     return todo, done
 
 
-def write_cache(cache: Cache, sources: Iterable[Path], mode: FileMode) -> None:
+def write_cache(cache: Cache, sources: Iterable[Path], mode: Mode) -> None:
     """Update the cache file."""
     cache_file = get_cache_file(mode)
     try:
         CACHE_DIR.mkdir(parents=True, exist_ok=True)
         new_cache = {**cache, **{src.resolve(): get_cache_info(src) for src in sources}}
         with tempfile.NamedTemporaryFile(dir=str(cache_file.parent), delete=False) as f:
-            pickle.dump(new_cache, f, protocol=pickle.HIGHEST_PROTOCOL)
+            pickle.dump(new_cache, f, protocol=4)
         os.replace(f.name, cache_file)
     except OSError:
         pass
@@ -4099,5 +6406,32 @@ def patched_main() -> None:
     main()
 
 
+def fix_docstring(docstring: str, prefix: str) -> str:
+    # https://www.python.org/dev/peps/pep-0257/#handling-docstring-indentation
+    if not docstring:
+        return ""
+    # Convert tabs to spaces (following the normal Python rules)
+    # and split into a list of lines:
+    lines = docstring.expandtabs().splitlines()
+    # Determine minimum indentation (first line doesn't count):
+    indent = sys.maxsize
+    for line in lines[1:]:
+        stripped = line.lstrip()
+        if stripped:
+            indent = min(indent, len(line) - len(stripped))
+    # Remove indentation (first line is special):
+    trimmed = [lines[0].strip()]
+    if indent < sys.maxsize:
+        last_line_idx = len(lines) - 2
+        for i, line in enumerate(lines[1:]):
+            stripped_line = line[indent:].rstrip()
+            if stripped_line or i == last_line_idx:
+                trimmed.append(prefix + stripped_line)
+            else:
+                trimmed.append("")
+    # Return a single string:
+    return "\n".join(trimmed)
+
+
 if __name__ == "__main__":
     patched_main()