When a process dies due to an exception, we don't see anything on stdout or stderr at all. This makes debugging very difficult - typically the result is a deadlock. The remote framework used to issue a SYS message to the logger, which was much more helpful. I've been doing a lot of printf debugging lately that would have been helped by some built-in error reporting.

I know there are plans to move the logger out of the core, which is good, but we should think about what to do here.