www.neuron.yale.edu

That is an oversight. cvode wraps ida and the former is persistent while the latter is destroyed/created whenever the model topology, or if the model requires a differential algebraic solver, changes. Extacellular requires ida. I forgot to make all the ida calls to reflect the existing cvode style f...

Ok. so there is no problem with command line tools and readline. I think the next diagnostic test is to install a distributed pkg file for NEURON and see if that works with your python3.7. Please try https://neuron.yale.edu/ftp/neuron/versions/alpha/nrn-7.6.2-18-gcfe5f00.x86_64-osx-py-37-36-27.pkg J...

-with-readline=no

I assume that is just a typo and you used --with-readline=no
Anyway, have you updated the command line tools. And perhaps you can reinstall python3.7.
I build the distribution with python3.7 and do not see this issue.

The icon launch mechanism is fixed with nrn-7.6.2-18-gcfe5f00.x86_64-osx-py-37-36-27.pkg in the alpha download folder. So all functionality on high sierra 10.13 exists with Mojave 10.14. Note that the first launch of an icon may pop up a system request for your permission. Mojave has increased the s...

you are missing a declaration of s
objref s

Please try
nrn-7.6.2-15-gbdf78f2+.x86_64-osx-py-37-36-27.pkg
from
https://neuron.yale.edu/ftp/neuron/versions/alpha/

The icons are non-functional in this version due to the missing launcher program. So you will have to use a terminal window for
nrnivmodl
nrngui
etc.

If installing from sources, use the configure option --with-readline=no Also a library I used for the icon launch mechanism has been deprecated out of existence. That can be temporarily worked around by modifying nrn/src/mac/Makefile.am to comment out a single line. ie. diff --git a/src/mac/Makefile...

Well, good. However I am truly astonished by this requirement. What specific anaconda version is this?

Ok. So the issue is related to anaconda python. turns out there are a number of google error references for vtable for std::__cxx11:: basic_stringbuf and I notice on that may be relevant. As it turns out,e GCC 5 introduced new implementations of std::string and std::list. You can try the workaround ...

I'm traveling and have a Mac powerbook with a virtual box guest ubuntu 16.4 with gcc 5.4.0. Are you using a tar.gz or the GitHub repository files? I would try the latter git clone http://github.com/neuronsimulator/nrn cd nrn sh build.sh ./configure ... etc. It may be worth while to focus on a stripp...

Shall I try installing NEURON again? I don't see much chance of success unless something substantive changes on your system. This is not a problem I have ever encountered. https://debian-administration.org/article/176/Reinstalling_packages_to_fix_problems What I would do next. is base="build-essent...

The openmpi is working. But I missed the fact in your earlier message that the problem has nothing to do with mpi as linking fails even in the absence of --with-paranrn. That makes me wonder if you have an incomplete g++ development environment or one that has inconsistent libraries. The build envir...

I found out that I had already openmpi.

Nothing wrong with that in principle. So for nrn
make distclean
./configure ... # just like you did before
make -j install

and does that succeed?

It is easier to see the relationship between stderr and stdout by making available the interleaved output in a single file via make >& temp But no need to do that in this case as it is undoubtedly similar to your first message. We already suspect that the problem is related to your mpich installatio...

do which mpicc which mpic++ which mpicxx which mpiexe all seem to come from the same place? The interviews installation seemed to work. But can you verify by launching $HOME/neuron/iv/x86_64/bin/idraw I'm going to guess that the mpich installation has a problem. Do things compile with make dist_clea...