The onexc argument was added two weeks ago to replace onerror: d51a6dc

Though it improves the third argument (replacing a 3-tuple with the exception object), I think we should also consider two further improvements while we're in the business of replacing onerror, as the opportunity may not come again for a long time:

1. The first argument, function, provides an unreasonably deep insight into the internals of rmtree(). The implementation already bends the truth, for example supplying os.path.islink when a junction check fails. Error reporting can already be achieved by printing or re-raising the exception. It's not used in the standard library (e.g. by tempfile). What legitimate use cases are there? I propose we remove it!
2. The second argument, path, is already be available via exc.filename. This is noted in os.walk()'s documentation of its onerror argument. Also propose we remove it.

If we make onexc accept a single argument, it would align with the onerror argument to os.walk() and os.fwalk(), and provide more relevant (/less redundant) information.