1- Avoid [type_term (eq_refl : @eq Type PROP PROP')] when [PROP] is not given. This has significant performance implications.
2- In th case PROP is given (i.e., only when the tactic is manually used), introduce all the foralls and lets.

This should fix iris-examples.

The types of propositions for monPred lemma need to be [monPred I PROP] and not [bi_car (monPredI I PROP)], otherwise iIntoValid fails in a very weird way. Seems to be related to a Coq bug.

BiIndexBottom class for bottom element in a bi index. monPred_all is a monoid morphism, and related big op lemmas.

Reorganize bi.monpred. Add unfolding manual lemmas for monPred_at, and use them in proofmode.monpred. Add big op lemmas for monpred_at.

Type classes for updates properties

See merge request FP/iris-coq!106

Special proof mode class for adding a modality to a goal

See merge request FP/iris-coq!107

This used to be done by using `ElimModal` in backwards direction. Having
a separate type class for this gets rid of some hacks:

- Both `Hint Mode`s in forward and backwards direction for `ElimModal`.
- Weird type class precedence hacks to make sure the right instance is picked.
  These were needed because using `ElimModal` in backwards direction caused
  ambiguity.