We may have been experiencing one of those bugs and calling our behavior good

I suppose it's also possible there's a regression, but we're looking into it now. Will let you know what we find

@ryanbrush: ^

We do not use insert unconditional or salience

We are using the deprecated accumulator function though.

Looks like #171 is what broke us.

We have a couple instances of the RHS using both a retract!, and an insert!

@devn Traveling right now so can't look in depth, but using retract with truth maintenance can create difficult to reason about situations in any rule engine. If there is some test case that does not behave as one would expect I'll take a look at it.

after thinking about it for a little bit, i see why this is broken.

clara is doing it's job correctly.