GitLab

between 1 and 2 Angstroms, or they fall outside the bounds of the neck lookup
tables.

At least on my Mac, this results in a C++ exception that does not get translated
into a Python exception (i.e., it is not catchable and crashes the application)

Check for nan positions was being done incorrectly

Fixed Fortran compilation errors caused by function names being too long

Fixgb

cutoff GB.

Grammar fixes in the doxygen comments.

Variable step size integrators were scaling the step size incorrectly

Make it so that random forces and integrators generate unique seeds each time a Context is made

Fix for #780

principle -> principal... I'm a bozo.

Catch nan coordinates before they can produce a segfault

random numbers.  Add 2 citations about Langevin synchronization.  Convert the
references.bib file to UNIX line endings.

Some windows changes

Fix typemap for str/bytes -> std::string

Fixed bug in reference CustomCompoundBondForce

stripUnits behavior with bytes

.message -> str()

to 0 here as well.