The main motivation of PEP 461 was to help ease migration from, and/or have a single code base with, Python 2. But bytes formatting don't support the %r code supported in Python 2. Instead it supports the %a code which is exactly equivalent to the %r code in Python 2 but doesn't supported in Python 2. So it is not so easy to migrate from or have a single code base with Python 2 it the code uses the %r opcode.

As far as bytes formatting supports the %s code (an alias to %b) purely for compatibility with Python 2, it would be good to support the %r as an alias to %a.

Yeah, we probably got carried away by purity concerns.

"As far as bytes formatting supports the %s code (an alias to %b) purely for compatibility with Python 2, it would be good to support the %r as an alias to %a."

I don't like this idea.

Sometimes practicality wins; it's why we allow %s, and we should also allow %r.

Both %s and %r need to be clearly documented as an aid to Py2/3 code bases, and not recommended for new code.

Serhiy, do you have time to take of this?

The implementation is easy -- just add "case 'r':" before "case 'a':". The hard (to me) part is the documentation, so that I'm out of the game.

New changeset 611fa301b807 by Ethan Furman in branch 'default':
Close issue23467: add %r compatibility to bytes and bytearray
https://hg.python.org/cpython/rev/611fa301b807

You should also update the PEP 461.

Added in
https://hg.python.org/peps/rev/7fe79194a4f2