[1844] in Release_7.7_team
minutes, 6/30
daemon@ATHENA.MIT.EDU (Dan Winship)
Wed Jun 30 14:16:25 1999
Date: Wed, 30 Jun 1999 14:16:12 -0400 (EDT)
Message-Id: <199906301816.OAA29373@antharia.mit.edu>
From: Dan Winship <danw@MIT.EDU>
To: release-team@MIT.EDU
Attending: danw, wdc, aurora, ajfox, miki, rbasch, ghudson, kcr,
jweiss, tb
1) SGI sucks again
We won't be getting a patch for nsd. We have to wait for the next
overlay, which they claim will not have too much unrelated crap in it.
2) BIND (also sucks again)
8.3 machines (running BIND 8.2) have been losing Hesiod since suomi
was updated to 8.2 (running BIND 8.1.2). There are various theories,
mostly involving TTLs, but no one is completely sure of anything.
ghudson: Our chances of fixing this are small. Should we back out to
BIND 8.1.2?
wdc: Have we sent a bug report to Vixie?
ghudson: We don't have anything coherent to report. I can try.
wdc: How much work will it be to back out?
ghudson: About an hour.
[Somewhere in here we agree to do it.]
danw: Do we want to back out suomi too since it will take a while to
get the patch release out?
ghudson: That's not the release-team's decision. And if the problem is
only with machines that have the bad data cached, it wouldn't help. So
I guess there's no need to do it.
3) 8.3 Public
Monday, July 12 - release-announce mail
Sunday, July 18 - release-announce reminder
Monday, July 19 (night) - Sun update
Tuesday, July 20 (~8pm) - Indy update
Wednesday, July 21 (~11pm) - O2 update
(Ops will do early Hesiod updates for Indy and O2, but not Sun.)
The O2 updated may be shifted around based on what happens with the
Indy update.
4) /bin/athena/bash -> /bin/sh symlink for 8.2?
We decided this wouldn't really help anyone, since there are more old
Linux and NetBSD boxes around than Athena workstations that won't be
going to 8.3.
5) Bug Tracking
ghudson: I looked at Bugzilla (Red Hat's bug-tracking system). It has
lots of dependencies, and uses lots of SQL. Basically no good for us.
Debian's is mail-server based. The server forwards things around,
developers just send mail. But it does lots of stuff with magic To:
addresses, so we'd need a dedicated virtual domain.
tb: They used to use X- headers instead of To: addresses, but it
caused problems and everyone liked the new system better. The To:
address is automatically propagated to replies, which the X- headers
aren't.
Discussion of bug tracking requirements is postponed. People agree
scarab should be resurrected.
