So essentially pushing the work to the TLB? Well, if there are people building moving GCs that manage to retain stable pointers that way, why not use it for a c

So essentially pushing the work to the TLB? Well, if there are people building moving GCs that manage to retain stable pointers that way, why not use it for a circular buffer as well. > The downside is that reallocation would be really slow, involving multiple syscalls, Scaling ring buffers up and down in size is not very performant anyhow, as a bunch of the elements in it tend to need to be copied.