This reverts commit 849abb8d, reversing changes
made to 8de0b894.  The merge was done
accidentally, before review happened.

This problem has been reported by Léon Gondelman.

Before, when using, for example wp_alloc, in an expression like:

  ref (ref v)

It would apply `tac_wp_alloc` to the outermost ref, after which it
fails to establish that the argument `ref v` is a value. In this
commit, other evaluation positions will be tried whenever it turn
out that the argument of the construct is not a value. The same
applies to store/cas/...

I have implemented this by making use of the new `IntoVal` class.

Instead of writing a separate tactic lemma for each pure reduction,
there is a single tactic lemma for performing all of them.

The instances of PureExec can be shared between WP tactics and, e.g.
symbolic execution in the ghost  threadpool

This patch was created using

  find -name *.v | xargs -L 1 awk -i inplace '{from = 0} /^From/{ from = 1; ever_from = 1} { if (from == 0 && seen == 0 && ever_from == 1) { print "Set Default Proof Using \"Type*\"."; seen = 1 } }1 '

and some minor manual editing

The WP construction now takes an invariant on states as a parameter
(part of the irisG class) and no longer builds in the authoritative
ownership of the entire state. When instantiating WP with a concrete
language on can choose its state invariant. For example, for heap_lang
we directly use `auth (gmap loc (frac * dec_agree val))`, and avoid
the indirection through invariants entirely.

As a result, we no longer have to carry `heap_ctx` around.

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

Robbert Krebbers authored 8 years ago and Ralf Jung committed 8 years ago

We do this by introducing a type class UpClose with notation ↑.

The reason for this change is as follows: since `nclose : namespace
→ coPset` is declared as a coercion, the notation `nclose N ⊆ E` was
pretty printed as `N ⊆ E`. However, `N ⊆ E` could not be typechecked
because type checking goes from left to right, and as such would look
for an instance `SubsetEq namespace`, which causes the right hand side
to be ill-typed.

Now we try to avoid adding them unnecessarily, so we don't have to remove them automatically any more.

And also rename the corresponding proof mode tactics.

rename program_logic.{ownership -> wsat}. It really is about world satisfaction and invariants more than about ownership.

As proposed by JH Jourdan in issue 34.

This is more consistent with CAS, which also can be used on any value.
Note that being able to (atomically) test for equality of any value and
being able to CAS on any value is not realistic. See the discussion at
https://gitlab.mpi-sws.org/FP/iris-coq/issues/26, and in particular JH
Jourdan's observation:

  I think indeed for heap_lang this is just too complicated.

  Anyway, the role of heap_lang is not to model any actual
  programming language, but rather to show that we can do proofs
  about certain programs. The fact that you can write unrealistic
  programs is not a problem, IMHO. The only thing which is important
  is that the program that we write are realistic (i.e., faithfully
  represents the algorithm we want to p

This commit is based on a commit by Zhen Zhang who generalized equality
to work on any literal (and not just integers).

This generalization is surprisingly easy in Iris 3.0, so I could not
resist not doing it :).

This makes stuff more uniform and also removes the need for the [inGFs]
type class. Instead, there is now a type class [subG Σ1 Σ2] which expresses
that a list of functors [Σ1] is contained in [Σ2].

Use it to prove that tests/barrier_client and tests/heap_lang are adequate.

This commit features:

- A simpler model. The recursive domain equation no longer involves a triple
  containing invariants, physical state and ghost state, but just ghost state.
  Invariants and physical state are encoded using (higher-order) ghost state.

- (Primitive) view shifts are formalized in the logic and all properties about
  it are proven in the logic instead of the model. Instead, the core logic
  features only a notion of raw view shifts which internalizing performing frame
  preserving updates.

- A better behaved notion of mask changing view shifts. In particular, we no
  longer have side-conditions on transitivity of view shifts, and we have a
  rule for introduction of mask changing view shifts |={E1,E2}=> P with
  E2 ⊆ E1 which allows to postpone performing a view shift.

- The weakest precondition connective is formalized in the logic using Banach's
  fixpoint. All properties about the connective are proven in the logic instead
  of directly in the model.

- Adequacy is proven in the logic and uses a primitive form of adequacy for
  uPred that only involves raw views shifts and laters.

Some remarks:

- I have removed binary view shifts. I did not see a way to describe all rules
  of the new mask changing view shifts using those.
- There is no longer the need for the notion of "frame shifting assertions" and
  these are thus removed. The rules for Hoare triples are thus also stated in
  terms of primitive view shifts.

TODO:

- Maybe rename primitive view shift into something more sensible
- Figure out a way to deal with closed proofs (see the commented out stuff in
  tests/heap_lang and tests/barrier_client).

The intropattern {H} also meant clear (both in ssreflect, and the logic
part of the introduction pattern).

Fixes issue #20.

Changes:
- We no longer have a different syntax for specializing a term H : P -★ Q whose
  range P or domain Q is persistent. There is just one syntax, and the system
  automatically determines whether either P or Q is persistent.
- While specializing a term, always modalities are automatically stripped. This
  gets rid of the specialization pattern !.
- Make the syntax of specialization patterns more consistent. The syntax for
  generating a goal is [goal_spec] where goal_spec is one of the following:

    H1 .. Hn : generate a goal using hypotheses H1 .. Hn
   -H1 .. Hn : generate a goal using all hypotheses but H1 .. Hn
           # : generate a goal for the premise in which all hypotheses can be
               used. This is only allowed when specializing H : P -★ Q where
               either P or Q is persistent.
           % : generate a goal for a pure premise.

And make constants P for which we do not want of_val P to reduce Opaque.

iSpecialize and iDestruct.

These tactics now all take an iTrm, which is a tuple consisting of
a.) a lemma or name of a hypotheses b.) arguments to instantiate
c.) a specialization pattern.

It is no longer triggered when posing [P ⊢ Q] with [P] an evar. This, for
example, makes sure that iApply pvs_intro works, which failed before.

That way, we do not have useless type annotations of the form
"v : language.val heap_lang" cluttering about any goal.

Note, that we could decide to eta expand everywhere (as we do for ∀
and ∃), and use the notation "WP e {{ Q }}" for "wp e ⊤ (λ _, Q)".