On Fri, Sep 25, 2020 at 6:06 PM Stefano Borini <stefano.borini@gmail.com> wrote:
Ricky, I'd love new dunders too but it would explode in complexity. The likelihood of being approved is pretty much non-existent as far as I understand.
If you want, we can write together a competing PEP with that solution, so that the Steering Council can have two ideas. I would not mind exploring that option a bit more, but it's likely to become an exercise.
You're kind to offer. I really don't feel qualified. I'm not even a professional developer. On Fri, Sep 25, 2020 at 10:47 PM Steven D'Aprano <steve@pearwood.info> wrote:
On Fri, Sep 25, 2020 at 05:05:52PM -0400, Ricky Teachey wrote:
* sheepishly, whispering mostly to myself *
All of that could be solved with new dunders.
What would be the signature of the new dunders?
Well, either new dunders, or the other approach where there is a flag the changes the way the index content are passed to the existing dunders, in this form: def __setindex__(self, value, *args, **kwds) But honestly I was apprehensive of even bringing it up again (hence my *sheepishly* comment) and now regret doing so. It's probably not worth rehashing it. Especially after Guido's summary post he provided-- when he looked at the various ideas and what it would take to implement them at the C level-- it feels like the idea was defeated. So I should probably just give over. However if any of the people who were in favor of the idea want to take up the flag for it, I would be very happy for them to do that. If it's just me at this point, I'm just not capable/qualified to make the case for it, so it isn't appropriate for me to continue hammering at it. Just bringing it up one last time to make sure nobody else is as passionate about it as me. --- Ricky. "I've never met a Kentucky man who wasn't either thinking about going home or actually going home." - Happy Chandler