Update Analysis Server highlight API and implementation.

Update Analysis Server highlight API and implementation.

I think this is a good enough compromise between the existing and IDEA granilarity.

Eclipse plugin CL will follow.

(IN)VALID_STRING_ESCAPE are not implemented yet.

R=brianwilkerson@google.com
BUG=

Committed: https://github.com/dart-lang/sdk/commit/739847402ba02e225117fefe8c2e5111bc75c255

[Brian Wilkerson, 2015-07-09 13:58:19]

brianwilkerson@google.com changed reviewers:
+ danrubel@google.com

[Brian Wilkerson, 2015-07-09 13:58:21]

This is a breaking change, so we need to update the version number in both the
spec and in our client code.

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/test/an...
File pkg/analysis_server/test/analysis/notification_highlights_test.dart
(right):

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/test/an...
pkg/analysis_server/test/analysis/notification_highlights_test.dart:599:
solo_test_LOCAL_FUNCTION() {
Remove "solo_" prefix.

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/tool/sp...
File pkg/analysis_server/tool/spec/codegen_java_types.dart (right):

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/tool/sp...
pkg/analysis_server/tool/spec/codegen_java_types.dart:134: writeln('package
com.google.dart.server.generated.types;');
I agree with Dan's original reason for changing this, and don't understand why
we'd change it back. That said, I'm OK with it if Dan is.

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/tool/sp...
File pkg/analysis_server/tool/spec/spec_input.html (right):

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/tool/sp...
pkg/analysis_server/tool/spec/spec_input.html:2537:
<value><code>UNTYPED_VARIABLE</code></value>
Do we make a distinction between explicitly and implicitly 'dynamic'? If not,
should we? And if not, I like "DYNAMIC_VARIABLE" better than "UNTYPED_VARIABLE".

[scheglov, 2015-07-09 15:36:23]

PTAL

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/test/an...
File pkg/analysis_server/test/analysis/notification_highlights_test.dart
(right):

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/test/an...
pkg/analysis_server/test/analysis/notification_highlights_test.dart:599:
solo_test_LOCAL_FUNCTION() {
On 2015/07/09 13:58:20, Brian Wilkerson wrote:
> Remove "solo_" prefix.

Done.

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/tool/sp...
File pkg/analysis_server/tool/spec/codegen_java_types.dart (right):

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/tool/sp...
pkg/analysis_server/tool/spec/codegen_java_types.dart:134: writeln('package
com.google.dart.server.generated.types;');
On 2015/07/09 13:58:20, Brian Wilkerson wrote:
> I agree with Dan's original reason for changing this, and don't understand why
> we'd change it back. That said, I'm OK with it if Dan is.

OK, I'm changing it back.
We need to update clients anyway.

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/tool/sp...
File pkg/analysis_server/tool/spec/spec_input.html (right):

https://codereview.chromium.org/1227143003/diff/1/pkg/analysis_server/tool/sp...
pkg/analysis_server/tool/spec/spec_input.html:2537:
<value><code>UNTYPED_VARIABLE</code></value>
On 2015/07/09 13:58:20, Brian Wilkerson wrote:
> Do we make a distinction between explicitly and implicitly 'dynamic'? If not,
> should we? And if not, I like "DYNAMIC_VARIABLE" better than
"UNTYPED_VARIABLE".

Thanks.
Changed in to DYNAMIC_VARIABLE and made a tweak for fields.

[Brian Wilkerson, 2015-07-09 16:16:13]

LGTM

https://codereview.chromium.org/1227143003/diff/20001/pkg/analysis_server/too...
File pkg/analysis_server/tool/spec/codegen_analysis_server.dart (right):

https://codereview.chromium.org/1227143003/diff/20001/pkg/analysis_server/too...
pkg/analysis_server/tool/spec/codegen_analysis_server.dart:41: writeln('import
com.google.dart.server.generated.types.*;');
Did this get missed, or is this change still needed?

[scheglov, 2015-07-09 17:09:29]

PTAL

UNRESOLVED_INSTANCE_MEMBER_REFERENCE is not implemented yet.
But as it is in API, we can implement it later.

[scheglov, 2015-07-09 17:09:31]

PTAL

UNRESOLVED_INSTANCE_MEMBER_REFERENCE is not implemented yet.
But as it is in API, we can implement it later.

[Brian Wilkerson, 2015-07-09 17:27:35]

You probably didn't want to ask for another review :-)

https://codereview.chromium.org/1227143003/diff/40001/pkg/analysis_server/too...
File pkg/analysis_server/tool/spec/spec_input.html (right):

https://codereview.chromium.org/1227143003/diff/40001/pkg/analysis_server/too...
pkg/analysis_server/tool/spec/spec_input.html:2508:
<value><code>INSTANCE_GETTER_DECLARATION</code></value>
I just noticed that there's no "[INSTANCE | STATIC | TOP_LEVEL]_[GETTER |
SETTER]_REFERENCE". Should there be? I assume that they're not all encoded as
*_FIELD_REFERENCE, because of explicit getters and setters. If we add them,
should we remove "*_FIELD_REFERENCE" (since technically you're never referencing
a field, but invoking the implicit getter/setter)?

[scheglov, 2015-07-09 18:19:49]

Committed patchset #3 (id:40001) manually as
739847402ba02e225117fefe8c2e5111bc75c255 (presubmit successful).