The following review has been posted through the commitfest application: make installcheck-world: tested, failed Implements feature: not tested Spec compliant: not tested Documentation: not tested
The argument for consistency with other functions that are variadic makes sense to me: although they are different from ordinary variadic functions in the type unification they do, their description in the manual simply calls them functions without calling attention to any way that they are different beasts. So, a reader might reasonably be surprised that VARIADIC doesn't work in the usual way.
This patch applies (with some offsets) but the build produces several incompatible pointer type assignment warnings, and fails on errors where fcinfo->arg is no longer a thing (so should be rebased over the variable-length function call args patch).
It does not yet add regression tests, or update the documentation in func.sgml.
