1. 17
  1.  

  2. 5

    Does anyone know the history of why only OpenBSD has BSD Auth?

    https://en.wikipedia.org/wiki/BSD_Authentication

    1. [Comment removed by author]

      1. 2

        think of bsdauth as a framework for auth microservices.

      2. 9

        We didn’t like PAM.

        1. 1

          Despite the name, BSD Auth isn’t part of the historical BSD codebase that Free/Open/NetBSD forked from, or else it probably would be found across the BSDs. It was a proprietary extension shipped only in BSD/OS, so none of the free-software BSDs had it until OpenBSD dug it up out of BSD/OS’s shallow grave.

          1. 2

            I may be mistaken about the timeline, but I think bsdi tried to give bsd auth to FreeBSD before openpam started. I’m not sure why it was rejected, before my time.

            1. 1

              Yeah, I’m curious why FreeBSD didn’t take it as well

              1. 1

                I did a bit of searching in the FreeBSD mailing list archives before posting that and didn’t really turn anything up. As far as I can tell there was never any kind of discussion about it either way, at least on the public mailing lists. So it’s possible that if BSDi did try to donate it to FreeBSD, it got lost in some bitbucket somewhere and never got serious consideration.

                The only message at all I can find trying to bring it up is this one from 2001 where someone noticed OpenBSD’s new support for it, and asks if FreeBSD should follow suit (but the query got not replies).

                1. 1

                  Right. I tried searching again, but didn’t find much. I suppose I could break down and ask people, but a public record would be preferable. Anyway, your comment about digging it out of the grave made me look because bsd/os wasn’t entirely dead at the time. :)

                  1. 1

                    For the record, I asked millert and he got the BSD auth code directly from the BSDi people.

                    1. 1

                      There’s a difference between mostly dead and all dead. All dead, is well, dead. Mostly dead means they still got something to live for, like selling their organs to the free software that will make them all dead.

                      1. 1

                        I also have false memories about such events, which means either that we have been blasted by the same orbital mind laser operated by BSDI, or that what we’re recalling is integration of login_cap(3). It seems like both originated in BSD/OS, as indicated by the HISTORY section of the man pages:

                        OpenBSD:

                        The login_getclass function first appeared in OpenBSD 2.8.
                        

                        NetBSD:

                        The login_getclass family of functions are largely based on the BSD/OS
                        implementation of same, and appeared in NetBSD 1.5 by kind permission.
                        

                        FreeBSD:

                        (tactfully staying silent on the matter, but code committed in early 1997)