With Coq 8.6, you can no longer have intro patterns that give more names than
the constructor has.  Also, patterns with too few names are now interpreted as
filling up with "?", rather than putting the unnamed parts into the goal again.

Furthermore, it seems the behavior of "simplify_eq/=" changed, I guess
hypotheses are considered in different order now.  I managed to work around
this, but it all seem kind of fragile.

The next compilation failure is an "Anyomaly: ... Please report", so that's what I will do.

This reverts commit 20b4ae55bdf00edb751ccdab3eb876cb9b13c99f, which does
not seem to work with Coq 8.5pl2 (I accidentally tested with 8.5pl1).

This makes type checking more directed, and somewhat more predictable.

On the downside, it makes it impossible to declare the singleton on
lists as an instance of SingletonM and the insert and alter operations
on functions as instances of Alter and Insert. However, these were not
used often anyway.

Initial commit by Amin Timany.

This seems to shorten type class search.

Also, slightly reorganize.

simplify_equality        => simplify_eq
simplify_equality'       => simplify_eq/=
simplify_map_equality    => simplify_map_eq
simplify_map_equality'   => simplify_map_eq/=
simplify_option_equality => simplify_option_eq
simplify_list_equality   => simplify_list_eq
f_equal'                 => f_equal/=

The /= suffixes (meaning: do simpl) are inspired by ssreflect.

Also, make our redefinition of done more robust under different
orders of Importing modules.

Since Coq 8.4 did not backtrack on eauto premises, we used to ensure
that hints like

  Hint Extern 0 (?x ≡{_}≡ ?y) => reflexivity.

were not used for goals involving evars by writing ?x ≡{_}≡ ?y instead
of _ ≡{_}≡ _.

This seems to be a legacy issue that no longer applies to Coq 8.5, so
I have removed these restrictions making these hints thus more powerful.

Also do some minor clean up.

And use more uniform variable names.

The port makes the following notable changes:

* The carrier types of separation algebras and integer environments are no
  longer in Set. Now they have a type at a fixed type level above Set. This
  both works better in 8.5 and makes the formalization more general.
  I have tried putting them at polymorphic type levels, but that increased the
  compilation time by an order of magnitude.
* I am using a custom f_equal tactic written in Ltac to circumvent bug #4069.
  That bug has been fixed, so this custom tactic can be removed when the next
  beta of 8.5 is out.

Important changes in the core semantics:
* Types extended with function types. Since function types are a special kind
  of pointer types, types now have an additional mutual part called "ptr_type".
* Pointers extended with function pointers. Theses are just names that refer
  to an actual function in the function environment.
* Typing environments extended to assign argument and return types to function
  names. Before we used a separate environment for these, but since the
  argument and return types are already needed to type function pointers, this
  environment would appear in pretty much every typing judgment.

As a side-effect, the frontend has been rewritten entirely. The important
changes are:

* Type checking of expressions is more involved: there is a special kind of
  expression type corresponding to a function designator.
* To handle things like block scoped extern function, more state-fullness was
  needed. To prepare for future extensions, the entire frontend now uses a
  state monad.

addresses.

The operation "mem_force Γ m a" used to apply the identify function to
pricisely the object "a", even in case "a" is an "unsigned char" address
refering to an individual byte. This caused the ctree substructure of the
entire subobject to disappear and had the undesired effect that:

  mem_force Γ a m ⊑{Γ,true@Γm} m

failed to hold (i.e. unused reads cannot be removed).

The refinement relation on addresses allows union references to be refined:

  (β2 → β1) → RUnion i s β1 ⊆ RUnion i s β2

The result is that frozen values are below their unfrozen variant, which made
it possible to prove that constant propagation (see constant_propagation.v) can
be performed on the level of the memory model.

Now it only performs injection on hypotheses of the shape f .. = f ..