Implement rudimentary module linking.

Implement rudimentary module linking.

Constructs the (generally cyclic) graph of module instance objects
and populates their exports. Any exports other than nested modules
are currently set to 'undefined' (but already present as properties).

Details:
- Added new type JSModule for instance objects: a JSObject carrying a context.
- Statically allocate instance objects for all module literals (in parser 8-}).
- Extend interfaces to record and unify concrete instance objects,
  and to support iteration over members.
- Introduce new runtime function for pushing module contexts.
- Generate code for allocating, initializing, and setting module contexts,
  and for populating instance objects from module literals.
  Currently, all non-module exports are still initialized with 'undefined'.
- Module aliases are resolved statically, so no special code is required.
- Make sure that code containing module constructs is never optimized
  (macrofy AST node construction flag setting while we're at it).
- Add test case checking linkage.

Baseline: http://codereview.chromium.org/9722043/

R=svenpanne@chromium.org,mstarzinger@chromium.org
BUG=
TEST=

Committed: https://code.google.com/p/v8/source/detail?r=11336

[Sven Panne, 2012-03-30 11:42:19]

LGTM when the nits are addressed.

https://chromiumcodereview.appspot.com/9844002/diff/1/src/full-codegen.cc
File src/full-codegen.cc (right):

https://chromiumcodereview.appspot.com/9844002/diff/1/src/full-codegen.cc#new...
src/full-codegen.cc:631: static_cast<PropertyAttributes>(READ_ONLY + DONT_DELETE
+ DONT_ENUM);
Style nit: In the rest of the code we use '|' instead of '+'.

https://chromiumcodereview.appspot.com/9844002/diff/1/src/full-codegen.cc#new...
src/full-codegen.cc:639: ASSERT(!value.is_null());
Do we really need this after the test in the previous statement?

https://chromiumcodereview.appspot.com/9844002/diff/1/src/heap.cc
File src/heap.cc (right):

https://chromiumcodereview.appspot.com/9844002/diff/1/src/heap.cc#newcode3835
src/heap.cc:3835: if (!maybe_map->To<Map>(&map)) return maybe_map;
No need for the template parameter, I think.

https://chromiumcodereview.appspot.com/9844002/diff/1/src/heap.cc#newcode3837
src/heap.cc:3837: MaybeObject* result = AllocateJSObjectFromMap(map, TENURED);
No need to name the result.

https://chromiumcodereview.appspot.com/9844002/diff/1/src/heap.cc#newcode4721
src/heap.cc:4721: if (!maybe_result->ToObject(&result)) return maybe_result;
Use templatized "To", removing the need for an explicit cast here.

[rossberg, 2012-04-10 14:01:15]

https://chromiumcodereview.appspot.com/9844002/diff/1/src/full-codegen.cc
File src/full-codegen.cc (right):

https://chromiumcodereview.appspot.com/9844002/diff/1/src/full-codegen.cc#new...
src/full-codegen.cc:631: static_cast<PropertyAttributes>(READ_ONLY + DONT_DELETE
+ DONT_ENUM);
On 2012/03/30 11:42:20, Sven Panne wrote:
> Style nit: In the rest of the code we use '|' instead of '+'.

Done.

https://chromiumcodereview.appspot.com/9844002/diff/1/src/full-codegen.cc#new...
src/full-codegen.cc:639: ASSERT(!value.is_null());
On 2012/03/30 11:42:20, Sven Panne wrote:
> Do we really need this after the test in the previous statement?

Well, the assertion is supposed to stay, while the test is only a temporary
hack. But actually, I removed the hack (here and in VisitModuleDeclaration for
all codegens) in favour of simply allocating an empty module in ParseUrlModule.

https://chromiumcodereview.appspot.com/9844002/diff/1/src/heap.cc
File src/heap.cc (right):

https://chromiumcodereview.appspot.com/9844002/diff/1/src/heap.cc#newcode3835
src/heap.cc:3835: if (!maybe_map->To<Map>(&map)) return maybe_map;
On 2012/03/30 11:42:20, Sven Panne wrote:
> No need for the template parameter, I think.

Done.

https://chromiumcodereview.appspot.com/9844002/diff/1/src/heap.cc#newcode3837
src/heap.cc:3837: MaybeObject* result = AllocateJSObjectFromMap(map, TENURED);
On 2012/03/30 11:42:20, Sven Panne wrote:
> No need to name the result.

Done.

https://chromiumcodereview.appspot.com/9844002/diff/1/src/heap.cc#newcode4721
src/heap.cc:4721: if (!maybe_result->ToObject(&result)) return maybe_result;
On 2012/03/30 11:42:20, Sven Panne wrote:
> Use templatized "To", removing the need for an explicit cast here.

Unfortunately, that doesn't work here, because the proper map is only set in the
next step. :(

Perhaps AllocateFixedArray & friends should take a map as argument, but that's a
refactoring we should do separately.