Issues installing legacy FEniCS on macOS Sonoma

Hello,
I’m installing Fenics from source on macOS Sonoma with an Intel Chip. We’re starting a project based on FEniCS.jl wrapper, which is why I require the legacy version.

I’ve tried installing from the Docker container without success following the instruction from the guide I get:

Unable to find image 'quay.io/fenicsproject/stable:current' locally
current: Pulling from fenicsproject/stable
docker: [DEPRECATION NOTICE] Docker Image Format v1 and Docker Image manifest version 2, schema 1 support is disabled by default and will be removed in an upcoming release. Suggest the author of quay.io/fenicsproject/stable:current to upgrade the image to the OCI Format or Docker Image manifest v2, schema 2. More information at https://docs.docker.com/go/deprecated-image-specs/.
See 'docker run --help'.

I’ve also tried using docker using this image. The container immediately crashes.

I’ve decided to build directly from source, and I’m able to install fenics-ffc.

import ffc
print(ffc.__version__)
2019.1.0.post0

I clone the corresponding dolphin version and doing cmake .. throws error messages when creating the demo files.

-- Generating form files in demo, test and bench directories. May take some time...
-- ----------------------------------------------------------------------------------------
CMake Error at CMakeLists.txt:626 (message):
  Generation of form files failed:

  Compiling 1 forms in demo/undocumented/multimesh-stokes/cpp...

  ffc -l dolfin -s -O -r auto MultiMeshStokes.ufl

  AttributeError: module 'numpy' has no attribute 'product'



-- Configuring incomplete, errors occurred!

I’m not sure what version of numpy to specify in my venv or what other flags should I indicate when calling cmake

I appreciate any guidance!

Hi.

Welcome to the FEniCS community.

Legacy dolfin is no longer recommended. If you necessary need it, see for instance Fenics in docker - #4 by dokken. There, you’ll note you can find the legacy dolfin docker image in:

Cheers.

1 Like

Hi Jesus,

Thanks so much for your response!
I understand dolfin is no longer recommended, but I was hoping there could be a work around at installing from source. I’m not sure if I can make it work having to install separate modules in different methods, i.e., I’m able to install some things from the repository directly.

I’m attempting to install through conda instead and hopefully that works.

Python env is very limited in macOS. Just in case, you can install the latest legacy dolfin installation (2019.2.0.dev) with anaconda:

conda install -c conda-forge -c conda-forge/label/fenics-dev fenics-dolfin mpich