On Thu, 3 Nov 2022 14:37:55 GMT, Jim Laskey <jlas...@openjdk.org> wrote:
>> src/java.base/share/classes/java/lang/template/ProcessorLinkage.java line 60: >> >>> 58: * @throws NullPointerException if any of the arguments are null >>> 59: */ >>> 60: MethodHandle linkage(List<String> fragments, MethodType type); >> >> I suggest changing the protocol here to be able to take all bootstrap >> arguments into account, and return a `CallSite` instead. That will allow a >> `ProcessorLinkage` to take the lookup and name into account as well, and >> allows returning e.g. a `MutableCallSite` as well. >> >> Maybe this can still be changed later as well though, since the interface is >> sealed. > > Yes - this will all go away during preview. I disagree with Jorn, - CallSite.dynamicInvoker() can be used to see a `MutableCallSite` as a `MethodHandle` so returning a `MethodHandle` is as powerful as returning a `CallSite`. - Having a `ProcessorLinkage` that takes a Lookup as parameter is a security risk because it means that a processor have full access to the user code that calls the processor at runtime. ------------- PR: https://git.openjdk.org/jdk/pull/10889
