The `bytearray()` value set to `bytes.partition()` and `bytes.rpartition()` is not converted to a `bytes()` value, keeping a `bytearray()` value in a tuple

Bug report

Bug description:

The bytearray() value set to bytes.partition() and bytes.rpartition() is not converted to a bytes() value, keeping a bytearray() value in a tuple as shown below:

v = bytes(b'ABCD')

                # ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓           # ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
print(v.partition(bytearray(b'BC')))  # (b'A', bytearray(b'BC'), b'D')
print(v.rpartition(bytearray(b'BC'))) # (b'A', bytearray(b'BC'), b'D')
                 # ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑          # ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑

The bytearray() value set to bytes.partition() and bytes.rpartition() should be converted to a bytes() value, not keeping a bytearray() value in a tuple because they are bytes functions and because of readability and consistency as shown below:

v = bytes(b'ABCD')

                # ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓           # ↓↓↓↓↓
print(v.partition(bytearray(b'BC')))  # (b'A', b'BC', b'D')
print(v.rpartition(bytearray(b'BC'))) # (b'A', b'BC', b'D')
                 # ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑          # ↑↑↑↑↑

In addition, the bytes() value set to bytearray.partition() and bytearray.rpartition() is converted to a bytearray() value, not keeping a bytes() value in a tuple as shown below:

v = bytearray(b'ABCD')

                # ↓↓↓↓↓↓↓↓↓↓↓↓                      # ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
print(v.partition(bytes(b'BC')))  # (bytearray(b'A'), bytearray(b'BC'), bytearray(b'D'))
print(v.rpartition(bytes(b'BC'))) # (bytearray(b'A'), bytearray(b'BC'), bytearray(b'D'))
                 # ↑↑↑↑↑↑↑↑↑↑↑↑                     # ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑

CPython versions tested on:

3.12

Operating systems tested on:

Windows

Exécutez d’abord les exemples Python fournis, puis localisez les implémentations de bytes.partition(), bytes.rpartition(), bytearray.partition() et bytearray.rpartition(), ainsi que leurs tests. La tâche est terminée lorsque les éléments du tuple renvoyé correspondent systématiquement au type de l’objet sur lequel le partitionnement est appelé, avec une couverture de régression pour les deux directions et les deux types d’arguments.