Lawrence Mitchell
Lawrence Mitchell
A single round of the cudf-python tests failed due to not raising pandas warnings where we expect to. But I would have thought this should also happen on different CTK...
Thanks @davidwendt, I think this is ready for another look when you have the time.
> Although you cannot really verify the results are _stable_, tests are helpful in case this diverges in the future and we want to make sure it doesn't just fail...
Possibly related to https://github.com/dask/distributed/pull/8314 ?
Not sure it will be, sorry. The approach I had there was not considered appropriate long term. I'll see if I can dig up the current state of any discussions
> Also the RMM Cython needs to be updated, and all uses of its current method. We decided that this didn't seem to be necessary, since, for now, the Cython...
This is a breaking change since any user of the `CallbackMemoryResource` must adapt their callback signatures. That said, I can't find any...
tl;dr: I think we should replace the pattern: ```cython cdef void c_foo(self) except * nogil: do_foo() def foo(self): with nogil: self.c_foo() ``` With ```cython def foo(self): with nogil: do_foo() ```...