On Thu, Jan 6, 2011 at 7:11 PM, Jonas Sicking <jo...@sicking.cc> wrote: > No, implementing the storage mutex is an orders of magnitude harder > than implementing a locking mechanism prevents the callback from > happening in multiple threads at the same time. I'm saying that with > my implementor hat on. They are two very different types of locks. A > major difference is that one is grabbed synchronously and the other > asynchronously.
Either way, if it satisfies the goal of having a strictly-correct spec and is enough to unblock access to Storage from workers, I'm fine with that (whether or not the locking is reliably implemented in practice). -- Glenn Maynard
