[2522] in Kerberos_V5_Development
Re: Current brokeness of source tree on where....
daemon@ATHENA.MIT.EDU (Tom Yu)
Fri Oct 3 02:43:55 1997
Date: Fri, 3 Oct 1997 02:43:41 -0400
To: epeisach@MIT.EDU
Cc: krbdev@MIT.EDU
From: Tom Yu <tlyu@MIT.EDU>
In-Reply-To: <9710010239.AA02701@kangaroo.mit.edu>
[changing the CC: line to krbdev rather than krbcore]
I've checked in a fix that *should* deal with the site.exp problem.
It fixes the quoting problems somewhat. Basically the new library
build system causes the generation of site.exp in tests/dejagnu to
fail due to some quoting nastiness, as KRB5_RUN_ENV now contains loads
more shell metacharaters than it used to. This has been worked around
in an interesting fashion.
For those not on krbcore where the discussion started, there was talk
of getsockname() failing with UDP sockets under Solaris 2.4 and 2.5,
and is reputed to be fixed in 2.6. This causes the udp sections of
the lib/rpc test to fail. I'm seriously considering punting that test
completely, as there is nothing in our tree that uses it. We could
leave the code there in the case that someone is demented enough to
try to use it. I'd personally rather see gssrpc rewritten, along with
the test suites....
On another note, there should be snapshots of the head revision of the
source tree nightly now. They'll be in
/mit/krb5/snapshot/krb5-current.*.tar.gz, or something like that.
I'll think about the possibility of getting those put on athena-dist
later.
---Tom
