Community
Participate
Working Groups
From bug 421543 commit 10: > > I also released an additional disabled test > > org.eclipse.jdt.core.tests.compiler.regression.InterfaceMethodsTest. > > _testBug421543b() > > that shows two problems. Bridge method is getting emitted in X it could be > > emitted in J for efficiency reasons. > > I'll check this later, need to align with Spec Part J which has > JVMS 4.9.2 > "Each invokespecial instruction must name ..., or a method in a direct > superinterface of the current class or interface."
(In reply to Stephan Herrmann from comment #0) > From bug 421543 commit 10: > > > > I also released an additional disabled test > > > org.eclipse.jdt.core.tests.compiler.regression.InterfaceMethodsTest. > > > _testBug421543b() > > > that shows two problems. Bridge method is getting emitted in X it could be > > > emitted in J for efficiency reasons. Should note that the test has been fixed and enabled via bug 421543. Only the potential performance issue remains.
(In reply to Stephan Herrmann from comment #1) > Should note that the test has been fixed and enabled via bug 421543. > Only the potential performance issue remains. In that event, retargetting to 4.4, so cycles leading to GA could be spent on more important tasks.
Turns out this is not just an optimization, but is also a correctness issue. See https://bugs.eclipse.org/bugs/show_bug.cgi?id=436350, I'll take over this one. *** This bug has been marked as a duplicate of bug 436350 ***
Verified for 4.5 M2 using build I20140915-2000