Kai Germaschewski

I'm not suggesting to change the overall design of the API. Variables would remain under IO, as they are now. I like the current API. However, there is really no...

If you're on linux, this might help: ``` [kaig1@login5 ~]$ findmnt --target $PWD TARGET SOURCE FSTYPE OPTIONS /autofs/nccs-svm1_home1 172.30.252.196:/nccs/home1 nfs rw,nosuid,relatime,vers=3,rsize=32768,wsize=32768 ``` (Note that findmnt has options to make it...

@FabienSalmon: I think what you may need is ```fortran call adios2_put(engine, varArray, single_tmp, adios2_mode_sync, ierr) ``` This may decrease performance, but it should make it work correctly.

I hit this problem, too, and decided to dig into it. The solution is pretty simple: ```diff diff --git a/CMakeLists.txt b/CMakeLists.txt index 995761ea..06240ec6 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -14,6...

I figured I'll bump this up, since the issue is still present. And I think I understand why it happens, but I'm not so sure what the right solution is....

So as I had a chance to look at this a little more, first of all the suggest partial fix above ``` - if (sizeof...(Args) >= this->dimension()) + if constexpr...

I think the reason is that `clib` provides functionality for one particular backend, the one that is generally used to generate the clib functionality, as that can not be templated...

Learned two things from the looking at the sycl CI: * We're still testing a function that I declared deprecated. * Apparently sycl doesn't like the ∂ symbol in the...

I guess I'll agree that while there was some intention behind the different size for multi-d index vs 1-d index, it's really not possible to not do this in a...

> Having both size_type and index_type is a bit confusing here. One is 1d, the other is N-d, but of course 1d is just a special case of N-d, so...