Install tests don't seem to use the environment provided by
`def setup_run_environment(…)`; we work around by providing the
necessary PYTHONPATH environment variable manually.

Until now we used a lot of try/except to keep going when preparing the
environment for the build; it should just work now, so let's fail early.

This should fix problems with finding system-installed headers instead of the
spack-provided ones (e.g. googletest). (In some cases sites might inject
include paths via compiler settings (or other means)).

We encountered build fails on some HPC sites due to openssl mismatches
between bare-metal tools and spack — especially curl and similar tools
for fetching.
Our build tool, waf, needs to find spack-provided libraries, e.g.
libclang.so, at runtime… so we will only set the environment variable
for this.

This one includes a fix in hxtorch that wasn't caught when testing
locally (at UHEI) against newer pybind11 versions.

This reverts merge request !195

This is a workaround for pysctrltp not detecting newer versions
of pybind11 correctly, and for `pyhxcomm` building a test binary linking
against stuff which requires linkage against libpython.
The real fix requires CR within the software stack and will only be
available later.

In spack 0.18 boost default libraries changed, see [0].

[0] https://github.com/spack/spack/blob/v0.18.1/var/spack/repos/builtin/packages/boost/package.py#L77

Suggested by CINECA.

This reverts commit 674e26c0

This reverts commit 017e5e23

Get rid of fxdiv, too…

Now we also modify the `SPACK_`-prefixed stuff for the spack compiler
wrappers.

This adds googletest's lib/lib64 isntall folder explicitly to the list
of libraries (it should have been found… but wasn't).

* a few commits in pynn-brainscales (2), grenade (2), halco (2), hxcomm
  (2), haldls (1), hwdb (1)
  * add interface-draft for plasticity rules
  * support for requestion server-side information
    (`get_bitfile_info` and `get_repo_state`)
* also cleans up the tag mess in the dependent repositories…
  really same tag everywhere
* same for oppulance; as software state didn't change, we don't require
  oppulance to match pynn-brainscales version (rc2 is sufficient)…
  relaxing constraint, rc2

This reverts merge request !105

Let's try to find all build problems in one try… keep building after
the initial error to find additional ones.

This should fix the error about an ambiguous template instantiation
for 'struct halco::common::detail::limits<
halco::hicann_dls::vx::v1::SynapseOnSynapseRow, void>' which isn't
triggered on newer g++s. We dropped support for old `v1` hardware; this
was planned anyways.

This reverts merge request !104