[lmi] depr.impldec

[Greg Chicares]

clang's '-Wdeprecated-copy-with-dtor' pointed out that we're relying
on a deprecated language feature [depr.impldec] in quite a few instances,
which I'd like to fix now, lest "deprecated" become "deleted".

Please review this branch:
  https://git.savannah.nongnu.org/cgit/lmi.git/log/?h=odd/rules_of_n
It passes every nychthemeral test, specifically including a complete
clang UBSan build. The strategy used was:
 - use the Rule of Zero where we can;
 - else, use the Rule of Five.

Somewhere, Scott Meyers said he disliked the Rule of Zero because he
sometimes experimentally adds a dtor as a convenient debugger breakpoint,
and adding one instead of five special members can break things. I don't
think I've used a debugger successfully in the present century, though.

I was torn between writing "= default" vs. "= delete".

I was tempted to remove "virtual" from dtors for classes that don't
manage resources (i.e., most lmi classes except any_member), but that
would add an extra sub-rule to the strategy:
 - when writing a non-virtual dtor for a polymorphic class, place it
   it a "protected:" section; and turn off the compiler warning that'll
   complain that the dtor isn't virtual
so I put that consideration aside in order to achieve the milestone,
because relying on a deprecated feature seems like a really bad idea.

Those are just my "known unknowns"; I'd be especially interested in
learning about any "unknown unknowns".