Support assignment of locals in analyzer2dart.

Support assignment of locals in analyzer2dart.

BUG=
R=paulberry@google.com

Committed: https://code.google.com/p/dart/source/detail?r=41352

[Johnni Winther, 2014-10-27 11:00:11]

johnniwinther@google.com changed reviewers:
+ paulberry@google.com, sigurdm@google.com

[Paul Berry, 2014-10-27 14:15:56]

lgtm

https://codereview.chromium.org/652403005/diff/1/pkg/analyzer2dart/lib/src/se...
File pkg/analyzer2dart/lib/src/semantic_visitor.dart (right):

https://codereview.chromium.org/652403005/diff/1/pkg/analyzer2dart/lib/src/se...
pkg/analyzer2dart/lib/src/semantic_visitor.dart:190: //
TODO(johnniwinther,paulberry): Is this a real thing?
This will only happen if the call

  node.leftHandSide.accept(ACCESS_SEMANTICS_VISITOR);

(below on line 238) returns null.  For an error-free program, I believe that can
only happen if the LHS of the assignment is an array indexing operation, i.e.:

  x[i] = y;

I think I'd prefer to handle this by either (a) adding a new AccessSemanticsKind
to represent the invocation of an index operator, or (b) simply handling it as a
special case in handleAssignmentExpression().

I'm not sure which of these two choices would be better.

[Johnni Winther, 2014-10-27 14:26:19]

https://codereview.chromium.org/652403005/diff/1/pkg/analyzer2dart/lib/src/se...
File pkg/analyzer2dart/lib/src/semantic_visitor.dart (right):

https://codereview.chromium.org/652403005/diff/1/pkg/analyzer2dart/lib/src/se...
pkg/analyzer2dart/lib/src/semantic_visitor.dart:190: //
TODO(johnniwinther,paulberry): Is this a real thing?
On 2014/10/27 14:15:56, Paul Berry wrote:
> This will only happen if the call
> 
>   node.leftHandSide.accept(ACCESS_SEMANTICS_VISITOR);
> 
> (below on line 238) returns null.  For an error-free program, I believe that
can
> only happen if the LHS of the assignment is an array indexing operation, i.e.:
> 
>   x[i] = y;
> 
> I think I'd prefer to handle this by either (a) adding a new
AccessSemanticsKind
> to represent the invocation of an index operator, or (b) simply handling it as
a
> special case in handleAssignmentExpression().
> 
> I'm not sure which of these two choices would be better.

Ahh. We need an AccessSemantics for index operations anyway. Adding a
TODO+giveup instead.

[Johnni Winther, 2014-10-28 09:06:44]

Committed patchset #2 (id:20001) manually as 41352 (presubmit successful).