Michael Schellenberger Costa
Michael Schellenberger Costa
@okomarov I would second that.
So does this also account for axis ticks and ticklabels?
Unfortunately that is not the whole truth. MSVC has checked iterators for their containers. Those e.g do additional bounds checking in Debug builds. The unwrapped iterators do not do anything...
That depends a lot on what you want to do. What VS version do you have, what CUDA version etc. (AFAIK 11.6 is needed for current MSVC STL)
@griwes I think I have addressed all comments, It would be great if you could just skimm over those changes
> @miscco I don't think you've actually pushed the changes? Oh I did push, just to a different repository 🤦
Thanks a lot, I have *no* idea why it sometimes creeps up into changes.
We passed the internal CI, and I pushed a final commit to _UGLIFY the macro definitions. I believe this is ready to be merged @griwes WDYT?
>2. Alternatively, we could move the emulation header into the libc++ layer and make the libc++ code paths also be fully backported this way. This is probably more work and...
> Hm, confirmed: https://godbolt.org/z/YPdYdaGTW > > @wmaxey is this a known problem? It looks like our structured binding tests for `tuple` are commented out: > > https://github.com/NVIDIA/libcudacxx/blob/4b1a1df666f3bac26f60cb06cda051ca6fc9b158/.upstream-tests/test/std/utilities/tuple/tuple.tuple/tuple.helper/tuple_size_structured_bindings.pass.cpp#L49-L67 > > The...