As it says on the title. This makes it hard to detect real kratos warnings in compilation output.
Apparently pybind developers have already solved it (https://github.com/pybind/pybind11/issues/1444) so I'd like to ask the @KratosMultiphysics/technical-committee permission to update the pybind version in this repo.
+1 on my side
Remember that we "patched" the pybind, @roigcarlo do you remember what we changed?
https://github.com/KratosMultiphysics/Kratos/commit/7960333336aa139ed227bc53fb9bd50f04c859d2 Ok it was this, take care
7960333 Ok it was this, take care
just for the record, I updated all of Kratos, so this hack should no longer be necessary
just for the record, I updated all of Kratos, so this hack should no longer be necessary
I'm not sure if all the non public ones have made the change.... @KratosMultiphysics/altair?
@pooyan-dadvand We do not have this change, in our core we still have:
using size_t = std::size_t;
Should we change this?
@ddiezrod I think that the important part is that, in the add_..._to_python files of your applications, you use namespace py = pybind11; instead of using namespace pybind11;. This was the source of the name clash with std::size_t.
Is this a change we can perform right now with our current Pybind version or it has to be done when pybind is updated?
It can be done right now.
@ddiezrod I think that the important part is that, in the
add_..._to_pythonfiles of your applications, you usenamespace py = pybind11;instead ofusing namespace pybind11;. This was the source of the name clash withstd::size_t.
It actually would be nice if you could do it soon. It is not much work, here is an example what is to be done (~10min/Application)
https://github.com/KratosMultiphysics/Kratos/pull/3270/files?utf8=%E2%9C%93&diff=split&w=1
@jcotela @philbucher Ok, we will take care of that. Thanks !
Most helpful comment
@jcotela @philbucher Ok, we will take care of that. Thanks !