Otherwise, ownership of cores in our ordered RA model will not be persistent.

Otherwise, whenever it cannot establish the Absorbing or Affine premise,
it will backtrack on the FromAssumption premise, causing a possible loop.
No idea why this happens, this may be a Coq bug...

As Aleš observed, in the ordered RA model it is not, unless the order on the
unit is timeless.

Thanks to discussions with Ales and Amin.

Coq also uses level 200 for these constructs.
Besides, heap_lang's match and if were also already at this level.

Now, associativity needs only to be established in case the elements are
valid and their compositions are defined. This is very much like the notion
of separation algebras I had in my PhD thesis (Def 4.2.1). The Dra to Ra
construction still easily works out.

Fix some longstanding renaming issues

See merge request FP/iris-coq!63

Replace/remove some occurences of `persistently` into `persistent` where the property instead of the modality is used.

Rename `UCMRA` → `Ucmra`
Rename `CMRA` → `Cmra`
Rename `OFE` → `Ofe` (`Ofe` was already used partially, but many occurences were missing)
Rename `STS` → `Sts`
Rename `DRA` → `Dra`

I have reimplemented the tactic for introduction of ∀s/pures using type
classes, which directly made it much more modular.

The advantage is that we can directly use a Coq introduction pattern
`cpat` to perform actions to the pure assertion. Before, this had
to be done in several steps:

  iDestruct ... as "[Htmp ...]"; iDestruct "Htmp" as %cpat.

That is, one had to introduce a temporary name.

I expect this to be quite useful in various developments as many of
e.g. our invariants are written as:

  ∃ x1 .. x2, ⌜ pure stuff ⌝ ∗ spacial stuff.