I know who Niko is. (And Mara too.)
Yes, I've read the whole thing. Including the parts that say unwinding isn't going away.
Note that I am not saying we should remove unwinding entirely: that’s not an option
And I also have enough experience to recognize any changes we may get -- like panic=abort having semantic impact -- are not likely to happen any time soon.
Yes, I saw your quote the first time, but as I said, it's out of context and incomplete, so a misleading quote at best.
The whole blog entry makes it quite obvious it'll be deprecated but that it could remain optional for parts of the code, at the cost of other features. The conclusion is clear:
I think the argument for deprecating unwinding boils down to this: unwinding purports to make cheap recovery tenable, but it’s not really reliable in the face of shared state. Meanwhile, it puts limits on what we can do in the language, ultimately decreasing reliability (because we can’t guarantee cleanup is done) and ease of use (borrow checker is stricter, APIs that would require cleanup can’t be written).
How could we deprecate it, though? It would basically become part of the ABI, much like C vs C-unwind. It’d be possible to opt-in on a finer-grained basis. In functions that are guaranteed not to have unwinding, the borrow checker could be more permissive, and must-move types could be supported.
I’m definitely tempted to sketch out what deprecating unwinding might look like in more detail. I’d be curious to hear from folks that rely on unwinding to better understand where it is useful— and if we can find alternatives that meet the need in a more narrowly tailored way!
To me it seems clear that this is very far from an actual proposal. Niko even says he's just "tempted" to sketch out what it might look like in practice, meaning he hasn't yet evaluated all the tradeoffs involved.
It doesn't look like something that will be done anytime soon, for sure. 
The point is that's something worth keeping in mind when advising about that feature. Seeing it's a real problem for the language's evolution, perhaps it's worth trying to find other solutions than relying on unwinding?
At least, that sounds safer to me and more constructive.
This topic was automatically closed 90 days after the last reply. We invite you to open a new topic if you have further questions or comments.