[1684] in Kerberos-V5-bugs

home help back first fref pref prev next nref lref last post

Re: appl/bsd/login.c smashes TZ environment variable

daemon@ATHENA.MIT.EDU (Marc Horowitz)
Sat Oct 21 21:59:13 1995

To: Sam Hartman <hartmans@MIT.EDU>
Cc: krb5-bugs@MIT.EDU, epeisach@MIT.EDU
In-Reply-To: Your message of "Sat, 21 Oct 1995 21:42:36 EDT."
             <199510220142.VAA14284@tertius.mit.edu> 
Date: Sat, 21 Oct 1995 21:59:03 EDT
From: Marc Horowitz <marc@MIT.EDU>

>> For example, consider getting initial tickets.  If a higher level
>> application wanted to use a simple interface, it would not be able to
>> deal with complex issues like determining how to get additional
>> information from a user for preauthentication.

In the most general case, you are right.  However, passing in a user
input callback function to a library function is a common thing to do.
It doesn't let you do arbitrarily complex user input (you probably
wouldn't define a retinal scan callback :-), but it does the job for
simple applications in 80% of the cases, where the preauth code just
needs the user to respond to a question (at a tty prompt, in a dialog
box, etc.).  For the other 20%, lower-level abstractions will need to
be used.

You can even provide a few prewritten callbacks for the common cases:
tty, Xaw, Motif, etc.  This makes it even easier for simple apps to
get complex things done easily.

FWIW, this is not an academic argument.  I did something similar to
this, but not quite as general, at OV, and it was immensely useful.

>> You can't make the abstractions simpler than the complexity mandated
>> by the problem you are solving.

No, you can't, but you can make the abstractions clever, and hide some
of the complexity even in cases where it doesn't seem to work at first
glance.

		Marc

home help back first fref pref prev next nref lref last post