Fix for issue 1317

[jeffriley]

Fix for issue 1317 - SN event not always logged to BSE log file when evolving MS merger products.

@ilyamandel I thought I had some questions about where we need to check for evolving MS merger products when there has been a stellar merger, but upon checking I think we're covered in all places, so the fix here should be all that's needed.

I am probably responsible for this problem manifesting - I have a vague recollection of adding a check for stellar merger towards the end of BaseBinaryStar::EvaluateBinary() thinking, at the time, "We don't need to do this if we've had a stellar merger". Well, we do if the stellar merger was a MS merger and we're evolving MS merger products... This is the check, now modified, at (now) line 2911 in BaseBinaryStar.cpp.

While investigating this problem I noticed that we don't always log a "TIMESTEP_COMPLETED" record to the BSE_Detailed_Output file for the very last timestep. We do log a "FINAL_STATE" record which has all the info required, but users (and particularly I think the detailed plotting code) look for the "TIMESTEP_COMPLETED" record type - so I changed the code to ensure a "TIMESTEP_COMPLETED" record is always logged, even if has the same information as the "FINAL_STATE" record.

[jeffriley]

Thanks @SimonStevenson - target branch is now correct :-)

[SimonStevenson]

Thanks for the fix Jeff! I have run the code and my example system is now correctly recorded in BSE_Supernovae. All good to go from me.