vendor decode_bytes (#249)

* add `decode_bytes` to `blackcompat`

* use `decode_bytes` from `blackcompat`

* correct the imports

Author: keewis

blackdoc/__main__.py

@@ -6,7 +6,7 @@
 from rich.text import Text
 
 from blackdoc import __version__, format_lines, formats
-from blackdoc.blackcompat import read_pyproject_toml
+from blackdoc.blackcompat import decode_bytes, read_pyproject_toml
 from blackdoc.colors import DiffHighlighter
 from blackdoc.console import err, out
 from blackdoc.diff import unified_diff
@@ -32,7 +32,7 @@ def check_format_names(string):
 def format_and_overwrite(path, mode):
     try:
         with open(path, mode="rb") as f:
-            content, encoding, newline = black.decode_bytes(f.read())
+            content, encoding, newline = decode_bytes(f.read(), mode)
 
         lines = content.split("\n")
 

blackdoc/blackcompat.py

@@ -3,12 +3,14 @@
 For the license, see /licenses/black
 """
 
+import io
 import os
 import sys
 from functools import lru_cache
 from pathlib import Path
 
 import tomli
+from black import Encoding, FileContent, Mode, NewLine, Preview, tokenize
 from pathspec import PathSpec
 
 
@@ -258,3 +260,38 @@ def gen_python_files(
             include_match = include.search(normalized_path) if include else True
             if include_match:
                 yield child
+
+
+def decode_bytes(src: bytes, mode: Mode) -> tuple[FileContent, Encoding, NewLine]:
+    """Return a tuple of (decoded_contents, encoding, newline).
+
+    `newline` is either CRLF or LF but `decoded_contents` is decoded with
+    universal newlines (i.e. only contains LF).
+    """
+    srcbuf = io.BytesIO(src)
+    encoding, lines = tokenize.detect_encoding(srcbuf.readline)
+    if not lines:
+        return "", encoding, "\n"
+
+    if Preview.normalize_cr_newlines in mode:
+        if lines[0][-2:] == b"\r\n":
+            if b"\r" in lines[0][:-2]:
+                newline = "\r"
+            else:
+                newline = "\r\n"
+        elif lines[0][-1:] == b"\n":
+            if b"\r" in lines[0][:-1]:
+                newline = "\r"
+            else:
+                newline = "\n"
+        else:
+            if b"\r" in lines[0]:
+                newline = "\r"
+            else:
+                newline = "\n"
+    else:
+        newline = "\r\n" if lines[0][-2:] == b"\r\n" else "\n"
+
+    srcbuf.seek(0)
+    with io.TextIOWrapper(srcbuf, encoding) as tiow:
+        return tiow.read(), encoding, newline