GitLab

pickleability.

- Add a docstring to element.Element.getByMass describing what it does
- Add a new element virtualsite with atomic number 0 (and alias it to extrapoint
  and lonepair)

constructor _should_ be able to take all of files in the CHARMM parameter set
distribution and correctly determine file type from just the name and/or
filename extension.

parameters to a ProteinStructure before calling createSystem.

Removed _d suffix from debug libraries

Fixed bug in OpenCL implementation of GBSA on CPU devices

really used CMake in-depth before).

Hopefully this fixes the Travis-CI build failures.

(again, code pulled from ParmEd) and update the tests to include this.

pick up some bugs that were not caught in my initial test following the port
from ParmEd.

All 4 tests pass right now, although no periodic systems are being tested.
Currently CHARMM, XPLOR, and VMD tests are all being run (for the same ala-3
system).

Disable implicit solvent tests, since there is no robust radii assignment
routine yet (the radii are stored in the Amber topology file).  I will port my
assignment routine from ParmEd here shortly and then re-enable the GB tests.

Make element basically immutable

correct list of modules (the existing list was for the ParmEd implementation
which separated out the OpenMM integration so not having it was not fatal).

Add a test case and example for running with CHARMM files.

Currently gives exactly the same energy as the same files passed through chamber
(which, in turn, was validated to machine precision against CHARMM itself).

Disable OpenCL platform on known-buggy macs

units on getPMEParameters

Change API reference links to 6.0

Fix broken Python wrappers