On Thu, Jan 9, 2020 at 1:47 AM language.lawyer--- via Std-Proposals <std-proposals@lists.isocpp.org> wrote:

I find it weird that storage of objects of volatile polymorphic class types can be touched when a virtual function is invoked, which happens in all "typical" implementations which use vtables and vptrs stored in objects to implement polymorphic behavior.
What if we just deprecate volatile-qualified polymorphic class types (as well as [[pointers|references] to|arrays of] them)?

Sounds rational, but I also don't really see the point. `volatile` is like `switch`: it sucks so much that nobody wants to touch it, not even to "fix" it. Especially since your proposed change sounds like a new special case, not a removal of special cases — your change would increase the length of the paper standard, not decrease it.

If you want to pursue it, I recommend talking to JF Bastien, who may or may not be interested in championing further "deprecating volatile" changes. After his recent deprecations that got into C++2a, maybe the idea has momentum, or maybe it's spent all its political capital, I'm not sure which.