pointer-constraints protocol: Removing lifetimes and persistency

[Drew DeVault]

On 2018-05-17  1:14 PM, Daniel Stone wrote:
> That's fine, but we would have to look at the trade-off: they would
> have to drop functionality they use, in exchange for something which
> is worse for them.

I guess I don't fully understand why it's worse for them. It should be
functionally equivalent, albiet with a frame or two of the pointer being
locked or unlocked at the incorrect time, which really doesn't matter.

Implementing persistent locks using only oneshots is basically a single
line of code afaict.

--
Drew DeVault