New boolean [empty] field in the eval_result record. It is set to [true] if a
result may represent an empty set.

Without using exceptions.

Avoids catching the exception in each case of the pattern matching,
and simplifies the implementation.

[Eva] Supports quantifier introducing mathematical variables

See merge request frama-c/frama-c!2521

- renames [eval_tlval] into [eval_term_as_lval], as it evaluates a term.
- renames [eval_thost_toffset] into [eval_tlval], as it evaluates a term_lval,
  and uses it instead of the previous [eval_tlval] when evaluating a term_lval.

[Eva] rename -eva-malloc-functions to -eva-alloc-wrappers and add calloc/realloc

See merge request frama-c/frama-c!2517

[Stdlib] Remove FCBuffer and FCSet

Closes #793

See merge request frama-c/frama-c!2522

Feature/db progress

Closes #790

See merge request frama-c/frama-c!2515

[typing] less warning on function applications

See merge request frama-c/frama-c!2519

[stdlib] Removes FCMap.
[eva] use `Map.union` instead of `Map.merge`

See merge request frama-c/frama-c!2520

[Libc] add spec for getsockname

See merge request frama-c/frama-c!2311

More specifically, do not warn if actual parameter has less `const` qualifiers
all along the type than formal (it was already handled for a `const` under
a single pointer, but not more). Of course, the reverse would still emit
a warning.

     Since Db.once are often registered in daemons

     It simplifies and is not really needed