Fix a bug where GWT-RPC would not generate some covariant array types, possibly causing inconsis...

Fix a bug where GWT-RPC would not generate some covariant array types, possibly
causing inconsistent serialization policies. Includes a test.

Fixes issue 7791

[stephenh]

http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
File
user/test/com/google/gwt/user/rebind/rpc/SerializableTypeOracleBuilderTest.java
(right):

http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
user/test/com/google/gwt/user/rebind/rpc/SerializableTypeOracleBuilderTest.java:2435:
private <T> String join(String delimiter, T... items) {
Why not reuse an existing join, like Guava's joiner?

http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
File
user/test/com/google/gwt/user/rebind/rpc/testcases/client/SubclassUsedInArray.java
(right):

http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
user/test/com/google/gwt/user/rebind/rpc/testcases/client/SubclassUsedInArray.java:30:
/** Array's leaf type. Not a root so its subtypes aren't cached. */
Why does the comment here say "Array's leaf type", but the next class is "class
LeafType"?

[skybrian]

http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
File
user/test/com/google/gwt/user/rebind/rpc/SerializableTypeOracleBuilderTest.java
(right):

http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
user/test/com/google/gwt/user/rebind/rpc/SerializableTypeOracleBuilderTest.java:2435:
private <T> String join(String delimiter, T... items) {
On 2012/11/21 15:39:58, stephenh wrote:
> Why not reuse an existing join, like Guava's joiner?
No good reason, just in a hurry.

http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
File
user/test/com/google/gwt/user/rebind/rpc/testcases/client/SubclassUsedInArray.java
(right):

http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
user/test/com/google/gwt/user/rebind/rpc/testcases/client/SubclassUsedInArray.java:30:
/** Array's leaf type. Not a root so its subtypes aren't cached. */
On 2012/11/21 15:39:58, stephenh wrote:
> Why does the comment here say "Array's leaf type", but the next class is
"class
> LeafType"?

Sorry about that! "Leaf" is used two different ways and I somehow didn't notice.
One is the type hierarchy: Base->Subtype->LeafType. But also, in the array code,
for a type such as Foo[][], we're calling "Foo" the leaf type.

We should probably decide on some consistent terminology and stick with it.

[cromwellian]

Usually in other part of the compiler, that is called 'ComponentType'
or 'ElementType'


On Wed, Nov 21, 2012 at 11:28 AM,  <skybrian@google.com> wrote:
>
>
http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
> File
>
user/test/com/google/gwt/user/rebind/rpc/SerializableTypeOracleBuilderTest.java
> (right):
>
>
http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
>
user/test/com/google/gwt/user/rebind/rpc/SerializableTypeOracleBuilderTest.java:2435:
> private <T> String join(String delimiter, T... items) {
> On 2012/11/21 15:39:58, stephenh wrote:
>>
>> Why not reuse an existing join, like Guava's joiner?
>
> No good reason, just in a hurry.
>
>
>
http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
> File
>
user/test/com/google/gwt/user/rebind/rpc/testcases/client/SubclassUsedInArray.java
> (right):
>
>
http://gwt-code-reviews.appspot.com/1869804/diff/1/user/test/com/google/gwt/u...
>
user/test/com/google/gwt/user/rebind/rpc/testcases/client/SubclassUsedInArray.java:30:
> /** Array's leaf type. Not a root so its subtypes aren't cached. */
> On 2012/11/21 15:39:58, stephenh wrote:
>>
>> Why does the comment here say "Array's leaf type", but the next class
>
> is "class
>>
>> LeafType"?
>
>
> Sorry about that! "Leaf" is used two different ways and I somehow didn't
> notice. One is the type hierarchy: Base->Subtype->LeafType. But also, in
> the array code, for a type such as Foo[][], we're calling "Foo" the leaf
> type.
>
> We should probably decide on some consistent terminology and stick with
> it.
>
>
> http://gwt-code-reviews.appspot.com/1869804/
>
> --
> http://groups.google.com/group/Google-Web-Toolkit-Contributors