So, so far, two 3rd party implementations who don’t want to see their code go into the stdlib, but who are in favour of someone else writing a stdlib implementation[1]? That’s perfectly fair, but still leaves us with the question of who would write (and maintain) the stdlib implementation.
This feels to me like a borderline case at the moment. No-one is particularly against the idea, but no-one is offering to do the work. I suspect that this will go nowhere unless/until someone writes an implementation.
Maybe at this point the next step should be to create a PR adding OrderedSet to collections. That might get some core dev attention and maybe a sponsor for a PEP.
-
Which presumably would mean they could recommend people use the stdlib version rather than theirs in future… ↩︎