p86create needs to handle unset IDs

[OPNA2608]

p86extract extracts all the entries with their IDs just fine but p86create doesn't care about where the user wants samples to be placed: It uses a simple for loop over all inputs & P86_AddSample to the next unset ID. And our test file happens to have unset IDs mixed with set IDs, so we can't compare hashes yet to check for regressions.