Feature or enhancement

Proposal:

This is a meta issue for this one pattern which is fairly small and well contained (ba = bytearray(), later does bytes(ba)). More complex refactors (ex. binacii.b2a_base64 usage) should get their own / separate issues.

Using .take_bytes() (gh-139871) can remove an allocation + copy in cases where a function builds up data in a temporary bytearray then at the end of the function returns bytes(ba). This can significantly speed up some code. For instance gh-141863 for asyncio.streams improved the asyncio_tcp pyperformance benchmark over 10%.

• base64 _b32encode, _b32decode
• wave _byteswap
• encodings.punycode
• encodings.idna
• urllib.parse unquote_to_bytes -- The bytearray.extend dramatically dominates time in profile, more complex refactor needed for measurable improvement.
• re._compiler charmap

Has this already been discussed elsewhere?

This is a minor feature, which does not need previous discussion elsewhere

Links to previous discussion of this feature:

No response

Linked PRs

• gh-141968: Use take_byes in base64 _b32encode and _b32decode #141971
• gh-141968: Use take_bytes in wave._byteswap #141973
• gh-141968: Use take_bytes in encodings.punycode #141974
• gh-141968: Use take_bytes in encodings.idna #141975
• gh-141968: Use take_bytes in re._compiler #141995
• gh-141968: Use take_bytes in zipfile._ZipDecrypter #142240
• gh-141968: Use take_bytes to remove copy in _pyio.BytesIO.read() #149850
• gh-141968: Use take_bytes to simplify and remove copy from pyrepl BaseEventQueue #149852
• gh-141968: Use bytearray.take_bytes() in more I/O paths #156989