MAINT: Amend byte cache

[j-t-1]

Cache string to byte conversions when the string length is greater than one.
#2726

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.14%. Comparing base (c96d01c) to head (e3136a6).
Report is 73 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2730   +/-   ##
=======================================
  Coverage   95.14%   95.14%           
=======================================
  Files          51       51           
  Lines        8547     8547           
  Branches     1703     1703           
=======================================
  Hits         8132     8132           
  Misses        261      261           
  Partials      154      154           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[stefan6419846]

I am not sure if this really makes sense or what the original intent of the old implementation has been. Until this is more clear (see my comment in #2726), I recommend against merging this.

[j-t-1]

@pytest.mark.parametrize(
    ("input_str", "expected"),
    [
        ("foo", b"foo"),
        ("😀", "😀".encode()),
        ("‰", "‰".encode()),
        ("▷", "▷".encode()),
        ("世", "世".encode()),
        # A multi-character string example with non-latin-1 characters:
        ("😀😃", "😀😃".encode()),
    ],
)
def test_b(input_str: str, expected: str):
    assert pypdf._utils.b_(input_str) == expected

Should this be:

def test_b(input_str: str, expected: bytes):
    assert pypdf._utils.b_(input_str) == expected

[stefan6419846]

Isn't your new code the same? If you are referring to the decorator - no, this cannot be removed. This defines the parameters/sub-tests to use for the test.

[j-t-1]

My comment could have been better. The change is to the type of the second argument from str to bytes (keeping the decorator).

[stefan6419846]

Yes, the expected parameter should be bytes, as we are encoding the strings here.

[j-t-1]

The caching of the existing function is maybe deliberate.