Richard S. Hall wrote:
Yes, this is a known issue. Other than using a timeout value so we can
fail if it starts to take a long time, there isn't much else that can
be done using the current algorithm.
In cases where there are many combinations of providers where some
lead to constraint violations, then the resolver can take a long time
because it is exhaustive.
We really need a better algorithm. I actually have a colleague who is
willing to look into the resolver algorithm, I just need to find the
time to sit down and discuss it with him.
-> richard
Just curious.... you are going to use a SAT solver?
Regards,
Alessandro Mottadelli
