Weird--I can't reproduce this problem. What is the output from conda info and conda list?

I think the problem is that the conda-forge channel is not enabled. The official PySCF conda packages depend on libstdc++ and libgcc >= 13.2.0, and those aren't available through Anaconda's main channel.

Try conda install -c pyscf -c conda-forge pyscf, or if that fails, conda create -n testenv123 -c pyscf -c conda-forge pyscf. You could also enable the conda-forge channel in your $HOME/.condarc.

Thanks for your help!

conda install -c pyscf -c conda-forge pyscf simply fails.

conda create -n xxx -c pyscf -c conda-forge python=3.9 pyscf works. But conda create -n xxx -c pyscf -c conda-forge python=3.9 then conda install pyscf -c pyscf fails. That doesn't matter... at least there's one approach works.

But, if the conda-forge channel has to be enabled, the correct instruction should be described in documentation, which is currently wrong. Or, just recommend conda install pyscf -c conda-forge since it's simple and it works.

By the way, may I ask why pyscf does not support installation with libgcc from default (main) channel (i.e. not packaged with libgcc from default channel) any more? What's the reason for binding to conda-forge channel?

I am one of the maintainer of the MOKIT(a MR workflow prog)'s conda package, which is built with gcc 11 from main channel (of course we could switch to conda-forge, but at the beginning I only knew the most common channel and use it -- as pyscf did two or three years ago). Some of our users reported that installing mokit is not compatible with pyscf from conda-forge and pyscf channel so I have to dig into it and ask the question above out of curiosity. Our current solution is to use pyscf from pip and it works good. And switching to conda-forge is another potential choice.

Sorry you are having problems with distributing MOKIT. If you have a moment, could you try running

conda create -n mokitenv -c mokit -c pyscf -c conda-forge --override-channels mokit pyscf

This worked for me. You can also add the options ... "libblas=*=*mkl" "python=3.11" to make NumPy use MKL and get a newer version of Python. Hopefully it works since it installs, but I didn't run any tests. In many cases, conda install will fail due to unsatisfiable constraints even though conda create succeeds.

Breakdown of the command, in case someone else finds this thread:

• conda create -n mokitenv -- create a conda environment named mokitenv
• -c mokit -c pyscf -c conda-forge --override-channels ... using the channels mokit, pyscf, and conda-forge, with priority in that order, and do not use any other channels
• mokit pyscf ... and install the packages mokit and pyscf.

For reference, here's the output of conda info on my computer (WSL2):

     active environment : None
            shell level : 0
       user config file : /home/chillenb/.condarc
 populated config files : /home/chillenb/miniforge3/.condarc
                          /home/chillenb/.condarc
          conda version : 24.1.2
    conda-build version : not installed
         python version : 3.12.2.final.0
                 solver : libmamba (default)
       virtual packages : __archspec=1=zen2
                          __conda=24.1.2=0
                          __cuda=12.4=0
                          __glibc=2.35=0
                          __linux=5.15.146.1=0
                          __unix=0=0
       base environment : /home/chillenb/miniforge3  (writable)
      conda av data dir : /home/chillenb/miniforge3/etc/conda
  conda av metadata url : None
           channel URLs : https://conda.anaconda.org/conda-forge/linux-64
                          https://conda.anaconda.org/conda-forge/noarch
          package cache : /home/chillenb/miniforge3/pkgs
                          /home/chillenb/.conda/pkgs
       envs directories : /home/chillenb/miniforge3/envs
                          /home/chillenb/.conda/envs
               platform : linux-64
             user-agent : conda/24.1.2 requests/2.31.0 CPython/3.12.2 Linux/5.15.146.1-microsoft-standard-WSL2 ubuntu/22.04.4 glibc/2.35 solver/libmamba conda-libmamba-solver/24.1.0 libmambapy/1.5.7
                UID:GID : 1000:1000
             netrc file : None
           offline mode : False

(also, recent conda versions have gotten much faster and it could be worth upgrading)

Sorry you are having problems with distributing MOKIT. If you have a moment, could you try running

conda create -n mokitenv -c mokit -c pyscf -c conda-forge --override-channels mokit pyscf

This worked for me. You can also add the options ... "libblas=*=*mkl" "python=3.11" to make NumPy use MKL and get a newer version of Python. Hopefully it works since it installs, but I didn't run any tests. In many cases, conda install will fail due to unsatisfiable constraints even though conda create succeeds.

Breakdown of the command, in case someone else finds this thread:

* `conda create -n mokitenv` -- create a conda environment named `mokitenv`

* `-c mokit -c pyscf -c conda-forge --override-channels` ... using the channels mokit, pyscf, and conda-forge, with priority in that order, and do not use any other channels

* `mokit pyscf` ... and install the packages mokit and pyscf.

For reference, here's the output of conda info on my computer (WSL2):

     active environment : None
            shell level : 0
       user config file : /home/chillenb/.condarc
 populated config files : /home/chillenb/miniforge3/.condarc
                          /home/chillenb/.condarc
          conda version : 24.1.2
    conda-build version : not installed
         python version : 3.12.2.final.0
                 solver : libmamba (default)
       virtual packages : __archspec=1=zen2
                          __conda=24.1.2=0
                          __cuda=12.4=0
                          __glibc=2.35=0
                          __linux=5.15.146.1=0
                          __unix=0=0
       base environment : /home/chillenb/miniforge3  (writable)
      conda av data dir : /home/chillenb/miniforge3/etc/conda
  conda av metadata url : None
           channel URLs : https://conda.anaconda.org/conda-forge/linux-64
                          https://conda.anaconda.org/conda-forge/noarch
          package cache : /home/chillenb/miniforge3/pkgs
                          /home/chillenb/.conda/pkgs
       envs directories : /home/chillenb/miniforge3/envs
                          /home/chillenb/.conda/envs
               platform : linux-64
             user-agent : conda/24.1.2 requests/2.31.0 CPython/3.12.2 Linux/5.15.146.1-microsoft-standard-WSL2 ubuntu/22.04.4 glibc/2.35 solver/libmamba conda-libmamba-solver/24.1.0 libmambapy/1.5.7
                UID:GID : 1000:1000
             netrc file : None
           offline mode : False

Thanks, I test this method in several machines and it works!