Add local type inference logic for integer literals.

Add local type inference logic for integer literals.

This CL introduces Fasta's first type inference rule, and hooks it up
to all the necessary testing infrastructure.  The rule is that a local
variable whose initializer is an integer literal should have an
inferred type of int.

Only variables occurring at top level within a method are inferred (we
don't yet recurse inside sub-blocks).

Additional type inference rules will be added in follow-up CLs.

R=ahe@google.com

Committed: https://github.com/dart-lang/sdk/commit/e8e8359cdf7039b562ea8c4541625f93db763a0e

[Paul Berry, 2017-04-19 14:28:55]

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

[ahe, 2017-04-20 08:36:42]

https://codereview.chromium.org/2828693003/diff/20001/pkg/front_end/lib/src/f...
File pkg/front_end/lib/src/fasta/target_implementation.dart (right):

https://codereview.chromium.org/2828693003/diff/20001/pkg/front_end/lib/src/f...
pkg/front_end/lib/src/fasta/target_implementation.dart:22: final bool
strongMode;
Why is this needed?

https://codereview.chromium.org/2828693003/diff/20001/pkg/front_end/testcases...
File pkg/front_end/testcases/inference/int_upwards_local.dart (right):

https://codereview.chromium.org/2828693003/diff/20001/pkg/front_end/testcases...
pkg/front_end/testcases/inference/int_upwards_local.dart:1:
/*@testedFeatures=inference*/
Missing copyright.

[Paul Berry, 2017-04-20 08:55:29]

https://codereview.chromium.org/2828693003/diff/20001/pkg/front_end/lib/src/f...
File pkg/front_end/lib/src/fasta/target_implementation.dart (right):

https://codereview.chromium.org/2828693003/diff/20001/pkg/front_end/lib/src/f...
pkg/front_end/lib/src/fasta/target_implementation.dart:22: final bool
strongMode;
On 2017/04/20 08:36:42, ahe wrote:
> Why is this needed?

Moved to KernelTarget, as we discussed.

https://codereview.chromium.org/2828693003/diff/20001/pkg/front_end/testcases...
File pkg/front_end/testcases/inference/int_upwards_local.dart (right):

https://codereview.chromium.org/2828693003/diff/20001/pkg/front_end/testcases...
pkg/front_end/testcases/inference/int_upwards_local.dart:1:
/*@testedFeatures=inference*/
On 2017/04/20 08:36:42, ahe wrote:
> Missing copyright.

Done.

[ahe, 2017-04-20 09:45:29]

lgtm

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/lib/src/f...
File pkg/front_end/lib/src/fasta/kernel/kernel_ast_factory.dart (right):

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/lib/src/f...
pkg/front_end/lib/src/fasta/kernel/kernel_ast_factory.dart:44: int charOffset =
TreeNode.noOffset,
Perhaps this should be a required parameter?

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/testcases...
File pkg/front_end/testcases/inference/int_upwards_local.dart (right):

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/testcases...
pkg/front_end/testcases/inference/int_upwards_local.dart:5:
/*@testedFeatures=inference*/
You should also be able to write this as:

/*@ testedFeatures = inference */

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/testcases...
File pkg/front_end/testcases/inference/int_upwards_local.dart.strong.expect
(right):

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/testcases...
pkg/front_end/testcases/inference/int_upwards_local.dart.strong.expect:6:
core::int x = 1;
This is kinda cool to see :-)

[Paul Berry, 2017-04-20 09:52:57]

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/lib/src/f...
File pkg/front_end/lib/src/fasta/kernel/kernel_ast_factory.dart (right):

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/lib/src/f...
pkg/front_end/lib/src/fasta/kernel/kernel_ast_factory.dart:44: int charOffset =
TreeNode.noOffset,
On 2017/04/20 09:45:29, ahe wrote:
> Perhaps this should be a required parameter?

I don't know--does it make sense for a variableDeclaration to always have a
charOffset?  It seems to me like it does, but there are two callers that don't
supply one (BodyBuilder.endFunctionName and BodyBuilder.pushNewLocalVariable),
so I'm not really sure what the story is there.

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/testcases...
File pkg/front_end/testcases/inference/int_upwards_local.dart (right):

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/testcases...
pkg/front_end/testcases/inference/int_upwards_local.dart:5:
/*@testedFeatures=inference*/
On 2017/04/20 09:45:29, ahe wrote:
> You should also be able to write this as:
> 
> /*@ testedFeatures = inference */

Acknowledged.

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/testcases...
File pkg/front_end/testcases/inference/int_upwards_local.dart.strong.expect
(right):

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/testcases...
pkg/front_end/testcases/inference/int_upwards_local.dart.strong.expect:6:
core::int x = 1;
On 2017/04/20 09:45:29, ahe wrote:
> This is kinda cool to see :-)

Agreed :)

[Paul Berry, 2017-04-20 10:36:24]

Description was changed from

==========
Add local type inference logic for integer literals.

This CL introduces Fasta's first type inference rule, and hooks it up
to all the necessary testing infrastructure.  The rule is that a local
variable whose initializer is an integer literal should have an
inferred type of int.

Only variables occurring at top level within a method are inferred (we
don't yet recurse inside sub-blocks).

Additional type inference rules will be added in follow-up CLs.
==========

to

==========
Add local type inference logic for integer literals.

This CL introduces Fasta's first type inference rule, and hooks it up
to all the necessary testing infrastructure.  The rule is that a local
variable whose initializer is an integer literal should have an
inferred type of int.

Only variables occurring at top level within a method are inferred (we
don't yet recurse inside sub-blocks).

Additional type inference rules will be added in follow-up CLs.

R=ahe@google.com

Committed:
https://github.com/dart-lang/sdk/commit/e8e8359cdf7039b562ea8c4541625f93db763a0e
==========

[Paul Berry, 2017-04-20 10:36:25]

Committed patchset #4 (id:60001) manually as
e8e8359cdf7039b562ea8c4541625f93db763a0e (presubmit successful).

[ahe, 2017-04-20 11:38:24]

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/lib/src/f...
File pkg/front_end/lib/src/fasta/kernel/kernel_ast_factory.dart (right):

https://codereview.chromium.org/2828693003/diff/60001/pkg/front_end/lib/src/f...
pkg/front_end/lib/src/fasta/kernel/kernel_ast_factory.dart:44: int charOffset =
TreeNode.noOffset,
On 2017/04/20 09:52:57, Paul Berry wrote:
> On 2017/04/20 09:45:29, ahe wrote:
> > Perhaps this should be a required parameter?
> 
> I don't know--does it make sense for a variableDeclaration to always have a
> charOffset?  It seems to me like it does, but there are two callers that don't
> supply one (BodyBuilder.endFunctionName and BodyBuilder.pushNewLocalVariable),
> so I'm not really sure what the story is there.

I think both of those are mistakes that would be caught by making this parameter
required :-)