Closed cbenzer closed 10 years ago
The argparse backport will probably be required, but I can fix the other problems. Thanks for the heads up!
Great. Looking forward to give the tool a test run, especially the restarting options as I need to break my simulations into smaller chunks.
I merged fixes for the issues you identified. With my python2.6.8 installation, everything is humming smoothly. Let me know if you have any issues.
Installed the updated code and first test run succeeded out of the box.
We are running all our scientific Linux boxes under CentOS 6 which means the default Python installation is 2.6. I ran into the following compatibility issues:
Issue 1 : ordereddict not available in Python 2.6. This could be resolved by installing the backport of ordereddict into Python 2.6 using "pip install ordereddict". Pip (python-pip) is available in EPEL repository. After that I had to change the line "from collections import OrderedDict" to "from ordereddict import OrderedDict" in the file openmm.
Issue 2: argparse not available in Python 2.6. This could be resolved by installing argparse into Python 2.6 using "pip install argparse".
After resolving issues 1 and 2 I was able to run "open-mm -h". Subsequently I tried running a simple test using the input.pdb file from the OpenMM 5.2 example folder.
Issue 3: Several warnings on TypeErrror:
This is probably related to this issue http://stackoverflow.com/questions/1713038/super-fails-with-error-typeerror-argument-1-must-be-type-not-classobj and is somehow masked by being default in Python 3.X but the need to be declared explicitly in version below 3.0
Issue 4: Not researched so far ... :-)
I would be grateful if you could work on the Python 2.6 compatibility and make life a little bit easier for all the RHEL and CentOS users.
Thanks a lot!