Review request for JDK-8013357: Javac accepts erroneous binary comparison operations

Alex Buckley alex.buckley at
Thu Jun 13 12:41:37 PDT 2013

With regard to in :

- In the second set of truth tables (for boxed primitive types), I like 
how the 3 moves across the second row :-)

- The comment "// Reference types" should be something like "// 
Important reference types", because the "// Boxed primitives" comment 
also deals with reference types.

- Line 167 says Int but means Integer.

- Line 271/272 should be "String, comparable to any reference type, 
excluding those convertible to a numeric types, but including any captures."

Other than that, good to go.


On 6/13/2013 12:28 PM, Eric McCorkle wrote:
> I've moved the new test into types as Jon suggested, and reordered the
> lists/truth tables as Alex suggested.  A new webrev is up here:
> On 06/11/13 12:31, Jonathan Gibbons wrote:
>> On 06/11/2013 06:10 AM, Eric McCorkle wrote:
>>> Good suggestion.
>>> Something else occurred to me: the scope of this test has grown beyond
>>> that of a simple regression test.  It should probably be moved to a
>>> subdirectory of test/tools/javac purposed for more exhaustive tests.
>>> Jon or Maurizio, do you have an suggestions?
>> Yes, test/tools/javac/types seems a better home.
>> We are trying to deprecate/minimize the use of bug numbers within
>> the test/tools/javac directory itself.
>> -- Jon

More information about the compiler-dev mailing list