Ready.
An E-net location, li may contain a single token, at a time. The "control state variables" of copycat82, correspond to E-net locations.
When E-nets are simulated, a transition waits, if the (selected) output location(s) for it, is/are full. To implement this policy with nondeterministic (no data, no time) reachability-test, simply pick, at each step, a transition that is not waiting. Let's refer to this as the WATE (WAit-Till-Empty) policy.
Petri nets are different, without any necessary bounds for tokens. A Petri net verifier would only list the cases that exceed a particular k-bound, and/or if unbounded. In any case, it is easy to write a Petri net verifer with a WATE policy, too.
A third alternative is to altogether dump away a net-model, stop the verifier, whenever some bound is exceeded.
copycat82 is vague about it. (See p.39, section 3.2.1) A location ("control state variable") is disabled if no tokens in it, enabled when with a single token, and otherwise it is considered a "design error." The concept of design-errorness was already pointed out in VD78 (p.188). VD78 does not modify the Petri net verifier in any way. VD78 strategy is inferrable as to point out, list, such cases. SARA CFA is similar. That is Petri nets, after all.
But those words in copycat82 may also mean to dump the faultily-built net away, or to adopt the E-net policy of WATE, i.e: do not trace the faulty-marked paths, which would lead to mutlple-tokens anywhere, any more.
In any case, if anything that does not fit E-nets, is considered a "design error," that means "the design" must be exactly similar to an E-net, as concerns the "control state variables" of copycat82, in correspondence to E-net locations.
The plagiarism of copycat82 is noticeable. The vagueness/absurdness of the problem to locate tokens in [varieties of] containers is only one of those self-contradictory consequences.