Ken Wesson wrote: > The interfaces are needed only if the function is passed around as an > object or dynamically rebound; if it's just called at a few fixed > sites, those calls can be directly to a known class.
The class names of function implementations seem to be a Clojure implementation detail. It has changed over the years, classes used to contain a unique numeric suffix. Closures don't have obvious names. If a function gets wrapped in something like a primitive-aware memoize, the client call sites shouldn't need to be recompiled. So I don't think a direct call to the implementation class is enough. > Even then, couldn't the compiler generate the needed interfaces for a > specific Clojure project on the fly along with the appropriate calls? I don't see why it couldn't. Actually I just hacked the Clojure compiler to do that, removed the limitation on 4 args, and removed all of the static IFn$*** classes, and it seemed to work fine. The static IFn$*** classes probably have some small value though. For instance, if some Java code is passed a dynamically created fn, which it knows to have a primitive interface, it can cast it to one of these pre-defined interfaces, whereas if the interfaces were only created dynamically, it wouldn't be easy to get the Java to compile in that case. 1.3 is currently an alpha, I suspect that this first stab at primitive args is currently deliberately conservative. -- Dave -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
