[14] in Hesiod
I am bewildered
daemon@ATHENA.MIT.EDU (Clyde Hoover)
Fri Nov 18 18:10:52 1988
From: clyde@mizar.cc.utexas.edu (Clyde Hoover)
To: hesiod@ATHENA.MIT.EDU
Cc: clyde@emx.utexas.edu
Date: Fri, 18 Nov 88 15:58:36 CST
I have been hacking upon Hesiod (from the first public release version)
for the last couple of days and I have a few questions/problems:
1. The examples show use of named data type 'TXT' but the Hesiod lib
routines ask for 'UNSPEC' records, never 'TXT' ones. Hence NO Hesiod
request will EVER be answered.
2. Named (with ALLOC_T_UNSPEC enabled) uses a strange format for
UNSPEC records (something about lengths and checksums), which is
totally undocumented (even looking at the source doesn't help too much),
but appears to be extremly clumsy for a human to produce.
3. The Hesiod lib routines totally ignore CNAME records but the examples
include them.
4. Attempting to run Hesiod outside of the normal named system is
a major pain. We already have BIND in our libc, but for testing I have
redirected Hesiod stuff to a different port, so I can't just link with
the standard BIND. There are name conflicts between libc and the
resolver code and the Hesiod code that caused no small end of grief.
I doubt that we will run Hesiod as a part of our standard nameserver,
so we will need to make using it 'standalone' easy to do. This release
will take some work to allow that to happen.
Finally, I don't see how the Hesiod distribution, as provided, can work
straight "out of the box." This can't be what Athena is using - it's a
mess that has taken me two days to figure out and test.
Could someone out there enlighten me about configuring a Hesiod system
and what all the various options to the libraries and named are and
which do I need to use?
-Clyde Hoover
