comportement was not very clear if an exception was raised.

TODO:

  - make the "on hover" prettyer (Find the size used by gtk for the
    popup message en set pp_set_margin to this size?)

  - Some transformations (ex: smoke_detector_*) should we mark them
    specialy during the registration i norder to warn the user?

 (metas, debug flags, transformations, formats) except for label.

This description is used in --list-*. The description can use any of
the formatting markup of Format "@ " "@[",...

Transformations can also specify from which metas and labels they
depend, and add informations about how they are interpreted.

TODO:
  - complete and correct the documentation
  - when a transformation use Trans.on_meta, it should be possible to
    add an interpretation of the metas in the documentation.
  - recover a summary version of --list-* ?
  - be able to export in latex?

  session pairing doesn't compute anymore the shape of the goal, it is
 done before. It was able to compute the shape only when the checksum
 of the task was different, but computing the checksum of the task is
 way more time consuming than computing the shape of the goal (and
 include it).

 So this commit simplify greatly the function and theoretically
 augment just a little the time spent. Experimentaly it's the inverse
 on max_matrix. Until "update_session: done" with or without modifying
 the checksums:

            before     |   after
without : 0.21-0.22 s  | 0.16-0.17 s
with    : 0.23-0.26 s  | 0.18-0.20 s

Otherwise, we would accept (mpv <- ghost v; let x = mpv in ...)
where both x and mpv are non-ghost.

(does not make much sense without polymorphism)

Coq realization for int.Power (mostly to keep Coq proofs that were in power.mlw)