Mark Harris

@mzient did you see my comment here: https://github.com/NVIDIA/libcudacxx/pull/158#discussion_r669279406 I ask because the code I mentioned still segfaults with your latest commits. ``` // segfault if (cuda::resource_view{nullptr} == cuda::resource_view{nullptr}) printf("equal\n"); ```...

I'm struggling with implementing `do_is_equal` in some derived classes. In many MR / adaptor implementations, there is a stored view of an upstream resource. So I may need to compare...

I still can't get some derived MRs to build due to `do_is_equal`. The problem is that in order to create a derived memory_resource type you have to implement ```bool do_is_equal(cuda::memory_resource...

After your latest changes this still doesn't compile: ``` cuda::stream_ordered_resource_view view{nullptr}; cuda::stream_ordered_memory_resource* mr{nullptr}; if (view == mr) { std::cout

The latter problem seems to be solved by adding overloads of `operator==` for `stream_ordered_memory_resource` like the ones you added for `memory_resource`. But adding these does not solve the problem with...

@mzient would you mind clicking "resolved" on the comment threads above that have been resolved?

@mzient you also should probably resolve the branch conflict mentioned below.

Is there a reason that `memory_resource` does not have `operator==`? When I have references to MRs, I have to create a view of at least one of them in order...

> > > Must you force push? As you can see, doing so disassociates the comment history from the code. > > > > > > Sorry... I did it...

Needed for RAPIDS libcudf (Python, Spark, and probably other clients)