Parsing empty bytes as HTML returns "b''"

Perhaps a bit of an edge case but if you try to parse b"" with the HTML parser, the returned text is "b''". With 4.12.3 the returned text is "".

#!/usr/bin/env python
import bs4
soup = bs4.BeautifulSoup(b"", "html.parser")
assert soup.get_text() == ""

Running this script with git-bisect:

5bf3787aa660ed87ecb40859639b3fa7e6497f5c is the first bad commit
commit 5bf3787aa660ed87ecb40859639b3fa7e6497f5c
Author: Leonard Richardson <email address hidden>
Date: Fri May 12 14:21:14 2023 -0400

    Went through dammit.py with mypy strict.

 bs4/dammit.py | 16 ++++++++--------
  1 file changed, 8 insertions(+), 8 deletions(-)

The line that broke this is in the __init__ of UnicodeDammit, line 800 at the time of writing on the master branch:

        # Short-circuit if the data is in Unicode to begin with.
        if isinstance(markup, str) or markup == b"":
            self.markup = markup
            self.unicode_markup = str(markup)
            self.original_encoding = None
            return

This used to read:

        if isinstance(markup, str) or markup == "":

Tested with Python 3.9 and 3.10
Neither html5lib nor lxml installed