Counting size of equations is wrong if tuples are involved

[Adrian Pop]

This model:

Real x[10];
equation
  (,,x[1],) = fcall(1); // wrongly counted as size 10!
  (,,x[2],) = fcall(2); // wrongly counted as size 10!
  ...
  (,,x[10],) = fcall(10); // wrongly counted as size 10!

So the system is not balanced as it has 10 variables and 100 equations.
This happens because we expand array crefs in algorithms to their full size and we also translate tuple = fcall() equations to algorithms.

This affects for example: ModelicaTest.Fluid.Dissipation.Verifications.HeatTransfer.Channel.kc_evenGapLaminar

[Adrian Pop]

I can fix this by sending a flag to lowerAlgorithm to know is an equation and not expand crefs in that case. Do you have any other (better way) to do this?

[Adrian Pop]

Fixed in r17557.