Release date: 2020-08-25
This release contains breaking changes. Be sure to follow migration steps before upgrading.
x.foo() in an oso policy could either be performing an attribute lookup or invoking a zero-arity method on
x. If looking up the
foo property returned a method, the host language libraries would transparently invoke it and return the result.
As of this release, parentheses are required for invocation.
x.foo performs a lookup, and
x.foo() invokes a zero-arity method.
oso can be run in debug mode by setting an environment variable
POLAR_LOG=1 This prints debug output when polar queries are evaluated to show what’s happening.
Other bugs & improvements¶
Improved performance of policies with many rules having ground (constant) parameters.
Improved performance of
inoperator (list membership) with many ground elements.
Stack traces return the original policy source instead of the internal version.
New FFI methods for passing print and warning messages from the core runtime to the language libraries.
Many thanks to Dan Callahan for the documentation suggestion!