[krbdev.mit.edu #8573] Re: [krbdev.mit.edu #8596] kinit segfault with Heimdal server

[Marc Dionne via RT]

On Wed, Jul 5, 2017 at 5:46 PM, Greg Hudson via RT
<rt-comment at krbdev.mit.edu> wrote:
> Nico reported this bug in April.  It was assigned #8573 (which I will
> merge this ticket into).  We fixed the encrypted challenge client code
> to check for no armor key on master, and marked the fix for pullup to
> the 1.15 and 1.14 release branches.  We haven't yet pulled up the fix to
> the release branches or issued patch releases, so it's not surprising
> that you would still see the bug in OS-packaged krb5 builds.

Sorry I didn't spot that one; but I didn't look at tickets in the
"review" state, and not sure that the subject would have stood out.

> I agree that it is also weird that the client can walk down the preauth
> mechanisms via prompter interrupts (or other prompter failures),
> especially for these two mechanisms.  Your patch might be a good idea
> independent of this segfault bug; I will think about it.

Seems to me that for a typical user not familiar with what's going on
under the hood, the intention when hitting ctrl-C is surely to abort
the whole command, not just the prompt and the single step associated
with it (these steps are not obvious/visible to the user).  So IMO it
makes sense to bail out here on ctrl-C.

Thanks,
Marc