I have taken a glance at the patch.They are similar. They both caused by a new version of VM.
But in bug 144742, It makes sense to lunch a Vm according the vm.name.
While to the library, i think it may encounter some issues. Because there exists a VM (like J9) support more than one kind of Class library ( J9 support IBM JDK classlib and Harmony classlib). And there also exits s classlib( like Harmony classlib ) support more than one kind of VM ( J9 & DRLVM )
So IMHO, i think we should not locate the class library just according to the java.vm.name property. Maybe the java.runtime.name or something else is a appropraite cadidate.