Re: heimdal on GNU HURD

[Thomas Bushnell, BSG]

"Jacques A. Vidrine" <n@nectar.com> writes:

> Are  referring to  the fact  that  I would  prefer to  use a  manifest
> constant versus  sysconf or looping until  a fit is found?   Clearly I
> don't think that is an inferior solution, but rather a practical one.

My point is that you are willing to deliberately introduce a bug!

Surely the C programming language has many fundamental flaws, and one
of them is that there is no convenient way to do this right.  But it
sounds like your preference is to use a manifest constant rather than
actually do the right thing!  (Where "right thing" means "work
properly under all the circumstances where it's possible to work
properly".)

> > You really think this is a serious cost in
> > a program that is doing lots of encryption??
> 
> No, I don't -- I just think it is a bug.  One which is easy to fix for
> the single-threaded case.

It's a bug to correctly support long hostnames?

> As I said in an earlier message,  I appreciate that.  I also happen to
> believe that it has been misapplied here.

I remember when passwords were limited to eight characters.  "What
possible value is there in a longer password?" we were asked.  We have
since learned that there is such value.

I don't know what possible value there is in huge hostnames.  But I
have enough humility to know that in nearly every case where there has
been a fixed limit on the length of a thing, it has turned out to be a
serious problem, and needed to be fixed.  I expect this case will be
the same.  

Since it's easy to simply handle the case correctly, why not do so?
Why deliberately cripple the software?

Thomas