Roman,

Thanks for letting us know. I have logged bug 6746196 to track this, and it will be fixed in a forthcoming JDK 7 build.

I agree that it would be good if javac followed the language specification (see bug 6400189). There might be a fear of causing programs that previously compiled to stop doing so, but in that case the right answer is probably a compiler flag, so that we could at least ask for strictness and avoid the conflicts with the Eclipse compiler.

Regards,
Éamonn McManus · JMX Spec Lead · http://weblogs.java.net/blog/emcmanus/


Roman Kennke wrote:
I think I posted some of this a while ago already, but it seems like
this kind of problems sneaks in every now and then. The JMX code has
some generics code that is not valid Java code but is accepted by javac
anyway (which is a bug in javac, which should be fixed IMO). Other (Java
compatible) compilers (like the Eclipse compiler) reject this code,
which is the correct thing to do. The attached patch fixes the (current)
problems in JMX.

/Roman

  

Reply via email to