There were exploits for all of these. Thus, they were not secure. There were no exploits for djbdns. What was your point again?
I don't know about my point, but I'm having trouble finding yours. You seem to be claiming, "There have not yet been any exploits found in djbdns, therefore it is secure."

Was OpenBSD secure until the moments a root exploit was publicized? Does the lack of knowledge of an exploit existing in an application somehow verify that application's security?

Why do bugs that have already been fixed somehow invalidate the security of a package in the future, compared to a package in which no bugs have happened to be found yet?

There's an awful lot of djb worshiping going on here. I seriously doubt djb happens to be the only living example in the world of a programmer who simply won't ever make a mistake. Very few people use djbdns compared to the uncountable masses that use popular open source server software. Don't you think there just might be a "more eyes spot more flies" corollary there?