This is a very hard bug to believe. And it was very hard
to reproduce, because it only bites under very particular
returns NEG, which is incorrect.
The *name* of the context is important. It must be the
same as the larger variable.
The context name may be quoted (context:'b) or not -- it
There must be TWO assumptions, with related
variables. But the order doesn't matter.
So apparently the context name is somehow getting
mixed up with the variables within the context.
Log in to post a comment.