Don't convert windows style \r\n to \n in file-contents-sorter

Instead preserve the original line endings.

With the `--unique` flag identical lines with different line endings
are collapsed to the first entry.

pre_commit_hooks/file_contents_sorter.py

@@ -22,6 +22,55 @@ PASS = 0
 FAIL = 1
 
 
+class Line:
+    """Wrapper to ignore end-of-line characters for sorting and comparison"""
+
+    def __init__(self, value: bytes, eol: bytes):
+        self._value = value
+        # Add an EOL if none present (can only happen to the last line)
+        if not self._value.endswith(b'\n'):
+            self._value += eol
+
+    def without_eol(self) -> bytes:
+        return self._value.rstrip(b'\n\r')
+
+    def unwrap(self) -> bytes:
+        return self._value
+
+    @classmethod
+    def key(
+        cls,
+        key: Callable[[bytes], Any] | None = None,
+    ) -> Callable[[Line], Any]:
+        if key is None:
+            return cls.without_eol
+        else:
+            def eol_key(val: Line) -> Any:
+                return key(val.without_eol())
+
+            return eol_key
+
+    def __eq__(self, o: object) -> bool:
+        if not isinstance(o, Line):
+            return NotImplemented
+        return self.without_eol().__eq__(o.without_eol())
+
+    def __hash__(self) -> int:
+        return self.without_eol().__hash__()
+
+
+def guess_eol(lines: list[bytes]) -> bytes:
+    if len(lines) == 0:
+        return b'\n'
+
+    for eol in [b'\r\n', b'\n']:
+        if lines[0].endswith(eol):
+            return eol
+
+    # Prefer '\n' if the first (only) line does not have a line ending
+    return b'\n'
+
+
 def sort_file_contents(
     f: IO[bytes],
     key: Callable[[bytes], Any] | None,
@@ -29,18 +78,16 @@ def sort_file_contents(
     unique: bool = False,
 ) -> int:
     before = list(f)
-    lines: Iterable[bytes] = (
-        line.rstrip(b'\n\r') for line in before if line.strip()
+    eol = guess_eol(before)
+    lines: Iterable[Line] = (
+        Line(line, eol) for line in before if line.strip()
     )
     if unique:
         lines = set(lines)
-    after = sorted(lines, key=key)
+    after = sorted(lines, key=Line.key(key))
 
     before_string = b''.join(before)
-    after_string = b'\n'.join(after)
-
-    if after_string:
-        after_string += b'\n'
+    after_string = b''.join(line.unwrap() for line in after)
 
     if before_string == after_string:
         return PASS

tests/file_contents_sorter_test.py

@@ -17,6 +17,7 @@ from pre_commit_hooks.file_contents_sorter import PASS
         (b'missing_newline', [], FAIL, b'missing_newline\n'),
         (b'newline\nmissing', [], FAIL, b'missing\nnewline\n'),
         (b'missing\nnewline', [], FAIL, b'missing\nnewline\n'),
+        (b'missing\r\nnewline', [], FAIL, b'missing\r\nnewline\r\n'),
         (b'alpha\nbeta\n', [], PASS, b'alpha\nbeta\n'),
         (b'beta\nalpha\n', [], FAIL, b'alpha\nbeta\n'),
         (b'C\nc\n', [], PASS, b'C\nc\n'),
@@ -67,6 +68,12 @@ from pre_commit_hooks.file_contents_sorter import PASS
             FAIL,
             b'Fie\nFoe\nfee\nfum\n',
         ),
+        (
+            b'Fie\r\nFie\nFoe\nfee\nfee\r\nfum\n',
+            ['--unique'],
+            FAIL,
+            b'Fie\r\nFoe\nfee\nfum\n',
+        ),
         (
             b'fee\nFie\nFoe\nfum\n',
             ['--unique', '--ignore-case'],
@@ -79,6 +86,24 @@ from pre_commit_hooks.file_contents_sorter import PASS
             FAIL,
             b'fee\nFie\nFoe\nfum\n',
         ),
+        (
+            b'linefeed\r\ncarriage_return\r\n',
+            [],
+            FAIL,
+            b'carriage_return\r\nlinefeed\r\n',
+        ),
+        (
+            b'carriage_return\r\nlinefeed\r\n',
+            [],
+            PASS,
+            b'carriage_return\r\nlinefeed\r\n',
+        ),
+        (
+            b'a\na\r\na\r\na\na\r\na\n',
+            [],
+            PASS,
+            b'a\na\r\na\r\na\na\r\na\n',
+        ),
     ),
 )
 def test_integration(input_s, argv, expected_retval, output, tmpdir):