Accuracy goal breaks MD5 sum

[kilian]

This input shows the problem:

process bug2_p = e1, E1 => e2, E2
compile
seed = 0
me = 0
mmu = 0
sqrts = 500
integrate (bug2_p) { accuracy_goal = 1 }
simulate (bug2_p) { n_events = 1 }

The simulate step fails due to a bogus report that the parameters don't match. Even without simulate, the grids won't be reread correctly when the same script is executed twice.

[Juergen Reuter]

Wrong priority order, after the build was cracked...

[kilian]

[We should have checked this earlier...]

[kilian]

The code above works now, r5862. Close the ticket when Brian's original code has been successfully tested.

[kilian]

OK, Brian's example now runs through once, and correctly reads and reuses the results in a second run. Closing.