This fixes issue #44.

    The idea on magic wand is to use it for curried lemmas and use ⊢ for uncurried lemmas.

This way we avoid the env_cbv tactic unfolding string related stuff
that appears in the goal and hypotheses of the proof mode.

The old choice for ★ was a arbitrary: the precedence of the ASCII asterisk *
was fixed at a wrong level in Coq, so we had to pick another symbol. The ★ was
a random choice from a unicode chart.

The new symbol ∗ (as proposed by David Swasey) corresponds better to
conventional practise and matches the symbol we use on paper.

Since env_cbv does not unfold these apps, we should do it ourselves,
to avoid ending up with partially evaluated terms.

This fixes issue 39.

There are now two proof mode tactics for dealing with modalities:

- `iModIntro` : introduction of a modality
- `iMod pm_trm as (x1 ... xn) "ipat"` : eliminate a modality

The behavior of these tactics can be controlled by instances of the `IntroModal`
and `ElimModal` type class. We have declared instances for later, except 0,
basic updates and fancy updates. The tactic `iMod` is flexible enough that it
can also eliminate an updates around a weakest pre, and so forth.

The corresponding introduction patterns of these tactics are `!>` and `>`.

These tactics replace the tactics `iUpdIntro`, `iUpd` and `iTimeless`.

Source of backwards incompatability: the introduction pattern `!>` is used for
introduction of arbitrary modalities. It used to introduce laters by stripping
of a later of each hypotheses.

And also rename the corresponding proof mode tactics.

Used in iRevert, iClear, iFrame, and for generalizing the IH in
iInduction and iLöb.

As proposed by JH Jourdan in issue 34.

Before, it failed when these tactics were invoked with persistent
hypotheses. The new behavior is more convenient when using these
tactics to build other tactics.

This comment mostly addresses issue #34.

There are still some issues:

- For iLöb we can write `iLöb (x1 .. xn) as "IH"` to revert x1 .. xn
  before performing Löb induction. An analogue notation for iInduction
  results in parsing conflicts.
- The names of the induction hypotheses in the Coq intro pattern are
  ignored. Instead, when using `iInduction x as pat "IH"` the induction
  hypotheses are given fresh names starting with "IH". The problem here
  is that the names in the introduction pattern are idents, whereas the
  induction hypotheses are inserted into the proof mode context, and thus
  need to have strings as names.

This closes issue 32.

This solves issue 33.

Before this commit, given "HP" : P and "H" : P -★ Q with Q persistent, one
could write:

  iSpecialize ("H" with "#HP")

to eliminate the wand in "H" while keeping the resource "HP". The lemma:

  own_valid : own γ x ⊢ ✓ x

was the prototypical example where this pattern (using the #) was used.

However, the pattern was too limited. For example, given "H" : P₁ -★ P₂ -★ Q",
one could not write iSpecialize ("H" with "#HP₁") because P₂ -★ Q is not
persistent, even when Q is.

So, instead, this commit introduces the following tactic:

  iSpecialize pm_trm as #

which allows one to eliminate implications and wands while being able to use
all hypotheses to prove the premises, as well as being able to use all
hypotheses to prove the resulting goal.

In the case of iDestruct, we now check whether all branches of the introduction
pattern start with an `#` (moving the hypothesis to the persistent context) or
`%` (moving the hypothesis to the pure Coq context). If this is the case, we
allow one to use all hypotheses for proving the premises, as well as for proving
the resulting goal.