Add parsing for tuple types

Add parsing for tuple types

This patch also:
-   fixes the use of predefined types as binding identifiers
-   moves the testgen mechanism in library file

This is based on https://codereview.chromium.org/1801633003

R=rossberg@chromium.org
BUG=
NOTRY=true
NOPRESUBMIT=true

[rossberg, 2016-03-17 16:51:38]

https://codereview.chromium.org/1810943002/diff/1/src/ast/ast.h
File src/ast/ast.h (right):

https://codereview.chromium.org/1810943002/diff/1/src/ast/ast.h#newcode2939
src/ast/ast.h:2939: class Type : public AstNode {
With all the new syntax it seems that this is now more than just "a type":
binder patterns are also represented as Type nodes, although they are not types
as such. Perhaps add a clarifying comment to that end.

https://codereview.chromium.org/1810943002/diff/1/src/parsing/parser-base.h
File src/parsing/parser-base.h (right):

https://codereview.chromium.org/1810943002/diff/1/src/parsing/parser-base.h#n...
src/parsing/parser-base.h:3310: trailing_comma = false;
Nit: you could make that into an else-branch for the following `if`.

https://codereview.chromium.org/1810943002/diff/1/test/cctest/test-parsing.cc
File test/cctest/test-parsing.cc (right):

https://codereview.chromium.org/1810943002/diff/1/test/cctest/test-parsing.cc...
test/cctest/test-parsing.cc:7532: "var x: [number[], (a: string, b?: number)]",
What about simpler cases like

[number, (a: string)]
[any, (a)]

https://codereview.chromium.org/1810943002/diff/1/test/mjsunit/harmony/typesy...
File test/mjsunit/harmony/typesystem/testgen.js (right):

https://codereview.chromium.org/1810943002/diff/1/test/mjsunit/harmony/typesy...
test/mjsunit/harmony/typesystem/testgen.js:26: class TestGen {
I assume this is unchanged?

https://codereview.chromium.org/1810943002/diff/1/test/mjsunit/harmony/typesy...
File test/mjsunit/harmony/typesystem/tuple-types.js (right):

https://codereview.chromium.org/1810943002/diff/1/test/mjsunit/harmony/typesy...
test/mjsunit/harmony/typesystem/tuple-types.js:26: function
ValidPrimaryTypes(size, proper=false) {
If I read the diffs correctly, then there is still a lot of duplication between
this new file and previously existing ones. Would it be easy to factor that out
somehow?

[nickie, 2016-03-18 11:12:01]

https://codereview.chromium.org/1810943002/diff/1/src/ast/ast.h
File src/ast/ast.h (right):

https://codereview.chromium.org/1810943002/diff/1/src/ast/ast.h#newcode2939
src/ast/ast.h:2939: class Type : public AstNode {
On 2016/03/17 16:51:37, rossberg wrote:
> With all the new syntax it seems that this is now more than just "a type":
> binder patterns are also represented as Type nodes, although they are not
types
> as such. Perhaps add a clarifying comment to that end.

Done.  I copied some information from the design doc as comments, both here and
in other classes that are used as covers.

https://codereview.chromium.org/1810943002/diff/1/src/parsing/parser-base.h
File src/parsing/parser-base.h (right):

https://codereview.chromium.org/1810943002/diff/1/src/parsing/parser-base.h#n...
src/parsing/parser-base.h:3310: trailing_comma = false;
On 2016/03/17 16:51:37, rossberg wrote:
> Nit: you could make that into an else-branch for the following `if`.

Done.

https://codereview.chromium.org/1810943002/diff/1/test/cctest/test-parsing.cc
File test/cctest/test-parsing.cc (right):

https://codereview.chromium.org/1810943002/diff/1/test/cctest/test-parsing.cc...
test/cctest/test-parsing.cc:7532: "var x: [number[], (a: string, b?: number)]",
On 2016/03/17 16:51:37, rossberg wrote:
> What about simpler cases like
> 
> [number, (a: string)]
> [any, (a)]

I'm adding the first.  The second is valid, 'a' being a type reference. 
(There's no symbol table for types yet.)

https://codereview.chromium.org/1810943002/diff/1/test/mjsunit/harmony/typesy...
File test/mjsunit/harmony/typesystem/testgen.js (right):

https://codereview.chromium.org/1810943002/diff/1/test/mjsunit/harmony/typesy...
test/mjsunit/harmony/typesystem/testgen.js:26: class TestGen {
On 2016/03/17 16:51:37, rossberg wrote:
> I assume this is unchanged?

Yes, just deleted from simple-types.js and brought here.

https://codereview.chromium.org/1810943002/diff/1/test/mjsunit/harmony/typesy...
File test/mjsunit/harmony/typesystem/tuple-types.js (right):

https://codereview.chromium.org/1810943002/diff/1/test/mjsunit/harmony/typesy...
test/mjsunit/harmony/typesystem/tuple-types.js:26: function
ValidPrimaryTypes(size, proper=false) {
On 2016/03/17 16:51:37, rossberg wrote:
> If I read the diffs correctly, then there is still a lot of duplication
between
> this new file and previously existing ones. Would it be easy to factor that
out
> somehow?

You're right.  Done.

[rossberg, 2016-04-07 14:52:24]

lgtm

[nickie, 2016-04-08 08:31:17]

Patchset #2 landed MANUALLY in experimental/nikolaos/types as:
bfc279ff613e859f2f2451e02aa6441e00c723ec
https://chromium.googlesource.com/v8/v8/+/bfc279ff613e859f2f2451e02aa6441e00c...

[nickie, 2016-04-08 08:31:33]

Description was changed from

==========
Add parsing for tuple types

This patch also:
-   fixes the use of predefined types as binding identifiers
-   moves the testgen mechanism in library file

This is based on https://codereview.chromium.org/1801633003

R=rossberg@chromium.org
BUG=
NOTRY=true
NOPRESUBMIT=true
==========

to

==========
Add parsing for tuple types

This patch also:
-   fixes the use of predefined types as binding identifiers
-   moves the testgen mechanism in library file

This is based on https://codereview.chromium.org/1801633003

R=rossberg@chromium.org
BUG=
NOTRY=true
NOPRESUBMIT=true
==========