GitLab

move old fnT+fnF versions of t_map, t_fold, etc. to a submodule

Rename as little as possible and keep the API.
Make all the necessary checks in Term and Decl.
Remove the duplicate code in Term but keep it elsewhere.
We will factorize the code as we go, without rush.

I love static typing!

programs: a label Init is automatically added at the beginning of each Hoare triple (we could discuss this choice); SF examples continued

The idea is that every big WP is built from "explainable" subformulas.
So, when we split for the first time, we stop at these subformulas.
The subsequent split will split them further.

Note that split_goal removes "stop_split" labels. Thus, if we have
an atomic assertion ("expl:precondition" "stop_split" true), then
the split_goal transformation will succeed and return a _different_
task, with the goal: ("expl:precondition" true). However, the second
split will return the same task exactly. Should we fix it?

At the moment, we lack explanation annotations over postconditions.

- the new syntax for localisation "labels" in Why is as follows:

    goal Toto #"file" line bchar echar#     - after an ident
    #"file" line bchar echar" (A and B)     - before a term/fmla

- the new syntax for buffer relocation is as follows:

    ##"file" line char##