Check Probability Consistency tool issues

[khoyme]

Within the Safety Tutorial on the Examples .git site, I am working with an instantiation of system implementation GPS.parts_SingleSensor

When I run the AADL->EMFTA tool, I get the fault tree with 4 elements OR’ed with the top level node having a probability of 0.0. This makes sense - in the example, the top node isn’t assigned a probability. If I run “Check Prob…” on the top level, I get an error of a mismatch. If I run “Generate Events Probability” the top level node gets a probability and the error goes away - they match. Again, this all makes sense.

Now, I added the following to the AADL model for GPS.parts_SingleSensor

properties
    emv2::OccurrenceDistribution => [ ProbabilityValue => 2.5e-5 ; Distribution => Poisson;] applies to FailStop;

This is the wrong probability. But when I re-instantiate, re-export to EMFTA, the top level node in the tree still has a probability of 0.0, not 2.5e-5.

Should the exporter be picking that new definition from the GPS.parts_SingleSensor implementation and populate that into the tree? (If it did, I would then expect that Check Probability Consistency would give an error as this definition doesn’t match the computation from the leaf nodes).