I feel like I'm missing something, because it seems like you are saying all of the following, some of which are contradictory. pkg(8) is like pkgin/nih pkg(8) replaces pkg_* pkg(8) is complementary to pkg_* and is optional pkg(8) needs some bootstrap mechanism (which pkgin doesn't really) In pkgsrc we seem to have evolved a situation where the basic package tools enable creating packages from source and adding and remove single packages (with simple fetch-and-install dependency management), and then there are two package managers that can be used on top of the basic tools. So certainly pkg(8) could be added as a third binary package manager, as a package in pkgsrc. But if the existing tools are going to stay (which seems clear, at least in the near term), then I don't understand why pkg(8) needs a bootstrap mechanism. There's a perhaps larger question, which is if there's a specification for both the package format itself and the control files stored in /var/db/pkg; I get the impression pkg(8) manipulates those directly and is thus a replacement for the union of pkg_* and pkgin/nih.
Attachment:
pgpyji3jyN_Mx.pgp
Description: PGP signature