GitLab

The underlying datastructure is not satisfying nor are the proofs but at
least there is a quick simple realization. It uses ClassicalEpsilon axioms
such as excluded middle, indefinite_description etc...
Some theory realizations are very far from reality because the axiom
characterization is small.

This patch also fully qualifies some realized types and constructors to
avoid inadvertent shadowing.

This removes lots of superfluous parentheses in the outer terms.

Since these files depend on Flocq, we are sure that Coq is at least 8.7.

This commit also removes some occurrences of 'fourier'.

Coq printer.
Also adding the "pretty-printed" generated realization.



N604-045 Coq code easier to read

* src/printer/coq.ml
Added indentation in Why3's Coq printer using the OCaml Format pretty
printing module.
Changed the display of logic formulas and terms only.

(cherry picked from commit 3453a65e0e5fca7e3aa16e22915ee3f079daf1c6)

Conflicts:
	src/printer/coq.ml
	src/transform/gnat_split_conj.ml

fix obsolete sessions

The former realization of bv used a bit of sign as the less significant
bit which should be the most significant. It did not raise problems
because to_int and to_uint were not linked by an axiom in the .why. Now,
it is the case, so we rewrote function twos_complement and made some
straightforward changes to the existing realization (surprisingly proofs
are the same).