On Sat, 10 Apr 1999, Nathan J. Williams wrote:

>  Perhaps worth noting is that compile_et and mk_cmds (and the
> libss and libcom_err stuff they come from) are not crypto-related at
> all - they're just utility libraries developed at MIT that krb4
> (and a few other packages) happens to use. So if one is willing to
> ignore the fact that, for NetBSD at the moment, they're only useful as
> tools for building krb4, they could be moved into the non-domestic
> part of the source tree and eliminate once and for all this particular
> bootstrapping problem.

Well, yes, the fact that they're only useful as tools for building krb4 is
why they don't belong in the main tree.  It especially makes sense to keep
them with krb4 if they are going to die along with that code, IMO.

--scott

PS - BTW, I suggested this same thing at least a year ago, but have since
been convinced that it's not a good idea.