[2050] in Kerberos-V5-bugs
Re: Beta 6 fails make check - SHS test 1 failed
daemon@ATHENA.MIT.EDU (Theodore Ts'o)
Wed Jun 26 21:57:27 1996
Date: Wed, 26 Jun 1996 21:57:29 -0400
From: "Theodore Ts'o" <tytso@rsts-11.mit.edu>
To: acheson@oak.cats.ohiou.edu
Cc: krb5-bugs@MIT.EDU
In-Reply-To: <SIMEON.9606261517.G@muahost.oak.cats.ohiou.edu> (message from
Todd Acheson on Wed, 26 Jun 1996 15:08:17 -0500)
From: Todd Acheson <acheson@oak.cats.ohiou.edu>
Date: Wed, 26 Jun 1996 15:08:17 -0500
On DEC UNIX 3.2c
The make executed cleanly, but make check fails:
"SHS test 1 failed"
I am currently looking at the code.
This is a known bug, which slipped out before we released the code. SHS
isn't used in the current krb5 release, so it isn't a disaster (and the
3DES/SHS code is *advertised* as going to change before the algorithms
settle down.)
The problem is that the init variable needs to be initialized to 1
before the test. It's pretty obvious bug in the
big-endian/little-endian detection code.
- Ted
