Install Python 2.7 32B on a anaconda 64B installation

[Piet]

On a Linux 64 BIT pc I installed conda 4.1.6 and upgraded to conda 4.1.8.
Then I installed into a new environment Python 2.7 32B

    bash-4.1$ set CONDA_FORCE_32BIT=1
    bash-4.1$ conda create -n py27_32 python=2.7
    Fetching package metadata .......
    Solving package specifications: ..........

    Package plan for installation in environment /scratch/vhelmont/anaconda/anaconda/envs/py27_32:

    The following NEW packages will be INSTALLED:

    openssl:    1.0.2h-1
    pip:        8.1.2-py27_0
    python:     2.7.12-1
    readline:   6.2-2
    setuptools: 23.0.0-py27_0
    sqlite:     3.13.0-0
    tk:         8.5.18-0
    wheel:      0.29.0-py27_0
    zlib:       1.2.8-3

    Proceed ([y]/n)? y

    Linking packages ...
    .....
    #
    # To activate this environment, use:
    # $ source activate py27_32
    #
    # To deactivate this environment, use:
    # $ source deactivate
    #
    bash-4.1$

 I activated it as follows:
    bash-4.1$ set CONDA_FORCE_32BIT=1
    bash-4.1$ source activate py27_32

My question now, how can I be sure I am using the 32 bit version of python 2.7. I have my doubts because when I checked the following command output:
   bash-4.1$ python
   >>> import platform                                                                    
  >>> platform.architecture()                                                            
  ('64bit', 'ELF')

it seems I am using the python 64 BIT version.

[Ilan Schnell]

The CONDA_FORCE_32BIT environment variable should never be used.  From the source code:
# CONDA_FORCE_32BIT should only be used when running conda-build (in order
# to build 32-bit packages on a 64-bit system).  We don't want to mention it
# in the documentation, because it can mess up a lot of things.

Just install the 32-bit Miniconda on your 64-bit Machine, instead.

Second, you can always be sure which which (32 or 64 bit) Python you are using by:

  * checking the ELF header of the binaries

  * looking at the pointer size of objects, i.e. tuple.__itemsize__
  * platform.architecture()

- Ilan

--
You received this message because you are subscribed to the Google Groups "conda - Public" group.
To unsubscribe from this group and stop receiving emails from it, send an email to conda+un...@continuum.io.
To post to this group, send email to co...@continuum.io.
Visit this group at https://groups.google.com/a/continuum.io/group/conda/.
To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/conda/5fe7b61d-b038-49c0-b5e0-810c185f1b7b%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

[Piet]

Hello,

Thanks for the info, installing the 32 bit version on a 64 bit station solved my problem.