Introduce Substitution class and Supertype class.

Introduce Substitution class and Supertype class.

Type annotations in the supertype clauses of a class are now Supertypes
instead of InterfaceTypes. In the current version, the two classes
contain the same information, but they are about to diverge in the
following ways:

- An InterfaceType may be nullable, whereas a supertype cannot.

- An InterfaceType may represent the exact class, a subclass, or
  a subtype of the given class.

- The type arguments to an interface type represent bounds, whereas
  the arguments to a supertype are always exact.

We also introduce a class Substitution that represents an operator
that replaces type parameters with types, depending on the variance
of their use site.

A substitution can be applied to a DartType or a Supertype, and can
be generated independently of how it will be applied.

BUG=
R=kmillikin@google.com

Committed: https://github.com/dart-lang/kernel/commit/e1c26f830e8cee1f73b8a0b5aa3dd665c0ce5da0

[asgerf, 2016-10-24 11:35:24]

asgerf@google.com changed reviewers:
+ kmillikin@google.com

[ahe, 2016-10-24 11:44:54]

ahe@google.com changed reviewers:
+ ahe@google.com

[ahe, 2016-10-24 11:44:54]

DBC: You may be looking for this terminology: a supertype is concrete, an
interface type is abstract. I have this terminology from Ole Agesen's thesis on
concrete type inference (page 4 in
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.93.4969&rep=rep1&typ...).

[Kevin Millikin (Google), 2016-10-24 13:35:48]

LGTM.

I agree with Peter that the terminology might need to be tweaked.  It's weird
that a Supertype is not just a supertype but also the mixin type and interface
types, so we might want a more generic term.

[ahe, 2016-10-24 13:42:12]

On 2016/10/24 13:35:48, Kevin Millikin (Google) wrote:
> LGTM.
> 
> I agree with Peter that the terminology might need to be tweaked.  It's weird
> that a Supertype is not just a supertype but also the mixin type and interface
> types, so we might want a more generic term.

I wasn't actually suggesting that Supertype isn't a good name. However, it felt
like the description of this CL could have benefited from the concrete/abstract
terminology.

I think Supertype is a great name for this particular concept.

[asgerf, 2016-10-24 14:05:50]

On 2016/10/24 13:42:12, ahe wrote:
> On 2016/10/24 13:35:48, Kevin Millikin (Google) wrote:
> > LGTM.
> > 
> > I agree with Peter that the terminology might need to be tweaked.  It's
weird
> > that a Supertype is not just a supertype but also the mixin type and
interface
> > types, so we might want a more generic term.
> 
> I wasn't actually suggesting that Supertype isn't a good name. However, it
felt
> like the description of this CL could have benefited from the
concrete/abstract
> terminology.
> 
> I think Supertype is a great name for this particular concept.

I looked at Agesen's termonology and I don't think it really applies here.

What he calls "interfaces" are structural types, "class types" are like our
InterfaceTypes, and "exact classes" are also going to be InterfaceTypes once
exactness is added to the type system.

The way I see it, Supertype annotations aren't really types, they are just part
of
a class definition.

[ahe, 2016-10-24 14:08:08]

On 2016/10/24 14:05:50, asgerf wrote:
> What he calls "interfaces" are structural types, "class types" are like our
> InterfaceTypes, and "exact classes" are also going to be InterfaceTypes once
> exactness is added to the type system.

That's because he's talking about Self. If you apply the same terminology to a
nominal type system, the terminology also works.

[asgerf, 2016-10-24 14:21:58]

Description was changed from

==========
Introduce Substitution class and Supertype class.

Type annotations in the supertype clauses of a class are now Supertypes
instead of InterfaceTypes. In the current version, the two classes
contain the same information, but they are about to diverge in the
following ways:

- An InterfaceType may be nullable, whereas a supertype cannot.

- An InterfaceType may represent the exact class, a subclass, or
  a subtype of the given class.

- The type arguments to an interface type represent bounds, whereas
  the arguments to a supertype are always exact.

We also introduce a class Substitution that represents an operator
that replaces type parameters with types, depending on the variance
of their use site.

A substitution can be applied to a DartType or a Supertype, and can
be generated independently of how it will be applied.

BUG=
==========

to

==========
Introduce Substitution class and Supertype class.

Type annotations in the supertype clauses of a class are now Supertypes
instead of InterfaceTypes. In the current version, the two classes
contain the same information, but they are about to diverge in the
following ways:

- An InterfaceType may be nullable, whereas a supertype cannot.

- An InterfaceType may represent the exact class, a subclass, or
  a subtype of the given class.

- The type arguments to an interface type represent bounds, whereas
  the arguments to a supertype are always exact.

We also introduce a class Substitution that represents an operator
that replaces type parameters with types, depending on the variance
of their use site.

A substitution can be applied to a DartType or a Supertype, and can
be generated independently of how it will be applied.

BUG=
R=kmillikin@google.com

Committed:
https://github.com/dart-lang/kernel/commit/e1c26f830e8cee1f73b8a0b5aa3dd665c0...
==========

[asgerf, 2016-10-24 14:22:00]

Committed patchset #1 (id:1) manually as
e1c26f830e8cee1f73b8a0b5aa3dd665c0ce5da0 (presubmit successful).