Until now, it was not possible to depend on the crate
nphysics without specifying at least one feature to select both the dimension (
2d) and the precision (
f64). This was confusing because failing to do so led to cryptic error messages. Users of
nphysics must now depend on either
nphysics3d. Those two crates will compile out-of-the box
and default to 32-bit floating point numbers. To use double precision, enable the feature
EDIT: after some clarifications about the working of cargo features, everything is parametrized by the scalar type now. Thus, starting with the version 0.2.0 of
nphysics’ crates most structures are generic (e.g. you have to use things like
World<f64>) and the
f64 feature has been removed.
I am not sure how this will affect compilation times for users so if the additional overhead is too dramatic, we might have to revert back to the old system (or another alternative on the futur).