Axel Huebl

Sounds ok. Labs that I am aware off using RHEL 6 will retire it by summer this year, yet `manylinux2010` is derived from Centos 6: https://www.python.org/dev/peps/pep-0571/ Anyway, manylinux2014 is around...

Simply adding it to the `lib/Makefile` with this patch ```patch From cd7d0165a0c547b7dee6387b03f3bb7b35937ad7 Mon Sep 17 00:00:00 2001 From: Axel Huebl Date: Wed, 14 Aug 2019 22:12:58 -0700 Subject: [PATCH] Makefile:...

problem occured downstream in @FrancescAlted et al.'s [c-blosc2](https://github.com/Blosc/c-blosc2) with an external Lizard build.

@inikep do you have hints what might be wrong with the install or usage of the decompress headers?

Thanks a lot, this fixes it for me! Just for packaging in Spack.io: you planning a bugfix release that will contain the fix? :)

> That's a lot of code (cost). > What's the benefit? I have been bitten by this in the past as well and the current inconsistency for 0-dim data (scalars)...

I did an extended test and the issue is present in 11.4 and newer: https://github.com/pybind/pybind11/issues/4193#issuecomment-1255646048

I reported the issue with the NVCC compiler team last week, but have not yet received a ticket number of tracking. Shall we try to find a work-around for the...

added in #4220

Let's do NVHPC (former PGI) compilers separately. They are a different stack.