[krbdev.mit.edu #626] shsUpdate, shsFinalize can trash core or produce bad results.

[Ken Raeburn via RT]

Hi Marcus.  I'm pretty sure we exchanged mail on this already, but I
didn't keep it, and Tom's being picky. :-)

Do we have your permission to include your test case t_shs3.c in our
distribution under MIT copyright?  Or at least under our usual licensing
terms?