[1059] in Kerberos_V5_Development
Re: [Doug Engert ] Krlogind and ss-962301
daemon@ATHENA.MIT.EDU (Sam Hartman)
Sun Apr 14 16:53:21 1996
To: Marc Horowitz <marc@MIT.EDU>
Cc: Sam Hartman <hartmans@MIT.EDU>, krbdev@MIT.EDU
From: Sam Hartman <hartmans@MIT.EDU>
Date: 14 Apr 1996 16:53:01 -0400
In-Reply-To: Marc Horowitz's message of Sun, 14 Apr 1996 16:34:43 EDT
>>>>> "Marc" == Marc Horowitz <marc@MIT.EDU> writes:
Marc> In message <tslzq8epopu.fsf@tertius.mit.edu>, Sam Hartman
Marc> <hartmans@MIT.EDU> writes:
>>> So, do I want to use streams or not with HP? It does work
>>> correctly now (see
>>> /mit/krb5/.build.sun4m_53/hp-build/appltelnet/telnetd/telnetd).
>>> I don't have krlogind working because Mark hasn't given us
>>> Marc's patches to krlogind.c so it doesnt compile.
Marc> hpux 9 or hpux 10? What I know for certain is that on the
Marc> hpux 10.x box I worked on, the hp vendor r* and telnet don't
Marc> use streams. Even if you've gotten them to seem to work,
Marc> I'm skeptical of using such a complex OS feature as streams
Marc> which basically isn't used by anything, and is therefore
Marc> probably poorly tested.
I'm somewhat reluctant to add additional OS-dependent
exceptions just to avoid using a complicated feature that is not
known to be broken. My experience is that streams have been
significantly more portable than other pty handling code and I try to
use it when present. I would rather wait until a problem is found in
the code.
I agree that more rigorous testing is appropriate and will
attempt to perform such testing.
--Sam
