Jonathan Halterman
Jonathan Halterman
Sure, which is why I'm hesitant to remove support for the current approach for Oracle/Open JDK users (most people) if not necessary.
How about an adaptive approach - attempt to load the class with Unsafe, if it fails then go through `SharedSecrets`? That would minimize inconvenience for existing users so they don't...
I would even keep the Unsafe approach for 10, 11 if it works, which for the Oracle and OpenJDKs I've tested with, it does.
Why do you think a multi-release jar would be necessary, and similar for the 1.8 requirement?
Hi Gian Maria - I don't have any major plans for TypeTools aside from continuing to maintain it to support newer and different JREs. So far I haven't heard from...
Would be happy to take a PR for this.
Thanks for digging into this. Care to submit a PR? :)
I'm guessing the impact of this is it causes the wrong method to be matched resulting in a failed type resolution? Happy to take a PR for this if there's...
The short answer is that this use case does not appear to be supported by the technique we're using to resolve lambda type information. The reason is, with: ``` Printer...
Hi @yaitskov, What's interesting about your scenario is that this works: ``` Consumer print = System.out::println; ``` But this doesn't: ``` Consumer print = System.out::println; ``` For whatever reason the...