GitLab

Adapted work on nearest_goal_right to be able to use it why3shell.
Now should be able to apply transformation and proof on current goal. Also
able to print the current tree and goal. Requires testing.

in one shell : ledit bin/why3shell file.xml

in another shell : tail -f why3shell.out

The sum of two positive (resp. negative) FP numbers should always be
positive (resp. negative), even when the result is zero.

This commit also cleans a bit some other parts of the specification of FP
operations.

We need to stop the transformation intro_premises to introduce variables
past a label. This allows us to keep variables in the goal (for counterex
generation) and be able to retrieve them as counterexamples.

* transform/intro_vc_vars_counterexmp.ml:
  changed vc_term_info so that it is not mutable anymore
  (do_intro): Removing the passing records to the do_intros calls which
may prevent us from seeing last vc_model
  (do_intro_vc_vars): adding a reference to keep the location of the vc

* transform/introduction.ml
  (intros): When encountering stop_intro label, the function should
stop introducing.

Add the possibility to keep currently unavailable theories in a session
during update. So the proofs are not lost. Currently not activated in
why3 command line tools.

  added in 9577acb9