Refactor capability code using fork.  Still too much refcounting, though.  Maybe this calls for a different design for pipelining...

Using promises for references to represent each branch of a fork doesn't work, because the fork hub will be deleted when the last branch resolves, before any continuation can run.  Instead, let's just institutionalize the concept of addRef() for owned pointers.

Interface code generation.  Also update method grammar to allow multiple returns, and allow specifying request/response structs instead of parameter lists.

Extend interface grammar to support inheritance, and update compiler to properly compile interfaces.

Get rid of no-longer-needed object-specific methods of DynamicStruct.  Use get().as<ObjectPointer>() instead.

Revamp generated interface to Object fields.  Now there is an ObjectPointer::{Reader,Builder}.  This will simplify the dynamic API (next change) and also makes it easier to delegate decisions about the object type to a function that doesn't know about the context where the object lives (i.e. by passing an ObjectPointer::Builder).

Factor Pointer{Reader,Builder} out from {Struct,List}{Reader,Builder}.  Will make it easier to deal with Object fields.

Let's go ahead and rename  to , acknowledging that this may expand to support pipeline operations that are not simply fetching members.

Improve and simplify event ordering.  When explicitly calling an EventLoop, events are now always added to the end of the queue, ensuring deterministic run order and making yield() obsolete.  OTOH, then() may queue events to run immediately, preempting other events.