I was wondering if fwrite_obj purposely overlooks null obj data or it should be checked. Phoneix Dracul had a core dump that pointed to this code being called on null obj_data, so it crashed his mud. The fix is simple.
The fix, in save.c fwrite_obj
if( iNest >= MAX_NEST )
bug( "fwrite_obj: iNest hit MAX_NEST %d", iNest );
if( !obj )
bug( "%s", "fwrite_obj: NULL obj" );
Just seems kinda obvious. I wonder why it was always overlooked?