UP012 fix / diff mode fails to remove unnecessary utf-8 (default) encoding parameter when the function call is split to multiple lines.

def ruff_up012_fix_buggy(a: int, b: str) -> bytes:
    return f"{a=} {b=}".encode(
        "utf-8",
    )

Using check --show-source:

$ ruff check --isolated --extend-select UP up012_fix_bug.py --show-source
up012_fix_bug.py:2:12: UP012 [*] Unnecessary call to `encode` as UTF-8
  |
2 |       return f"{a=} {b=}".encode(
  |  ____________^
3 | |         "utf-8",
4 | |     )
  | |_____^ UP012
  |
  = help: Remove unnecessary `encode`

Found 1 error.
[*] 1 potentially fixable with the --fix option.

Trying to --diff:

$ ruff check --isolated --extend-select UP up012_fix_bug.py --diff

error: Autofix introduced a syntax error. Reverting all changes.

This indicates a bug in `ruff`. If you could open an issue at: ...

I'm pretty sure ruff is leaving the trailing comma behind, leading to a syntax error.

Looks like using pyupgrade collapses the call to a single line.

When the original statement is on a single line everything works.

Note that the help message "Remove unnecessary encode" is misleading. This case is about removing the unnecessary utf-8 parameter to encode(), as that is the same as the default value. In case of f-strings there's no direct "bytes" variant (regular strings could be written as b"foobar" directly).