Huh, where are the accesses of free()d memory happening?
See below. But, it's now moot . . .
Oh, hm. But I am actually concerned about that. It looks like there's
no free_headercomps() function which would reset all of the folder_comp/
part_comp variables in mhshowsbr.c back to NULL.
And yes, the format engine's memory management is terrible; my plan
is to make it so you can allocate and manage multiple component tables
(right now there is only one component table permitted, so doing things
like this are hard).