[4639] in Release_7.7_team
Minutes of 2004-06-30 release team meeting
daemon@ATHENA.MIT.EDU (Greg Hudson)
Wed Jun 30 13:43:08 2004
Date: Wed, 30 Jun 2004 13:43:04 -0400
Message-Id: <200406301743.i5UHh4d6017385@equal-rites.mit.edu>
From: Greg Hudson <ghudson@MIT.EDU>
To: release-team@MIT.EDU
Attending: ghudson rbasch wdc
1. Public release
We will schedule the public release for July 12 and 13. There will be
one patch release before then.
2. Public workstation verification
We agreed that the best short-term fix is to allow manual interruption
of the verification by hitting return. Running verification in the
background could adversely affect machine performance and have other
problems.
In the long run, the answer is probably to switch from doing read
checks to doing stat checks; we could do this by using os-checkfiles
instead of rpm -V. Doing read checks likely isn't buying us anything;
casual modification of the workstation should be revealed by stat
checks, and careful deliberate modification is more easily
accomplished by disabling /etc/athena/verify than it is by resetting
the stat information. With stat checks, we probably don't have to
worry about decoupling verification into two pieces as mitchb
suggested in source-developers [604], because it would be fast enough
to run at boot time and at full reactivate time.
[More information revealed by post-meeting investigation:]
* There's no need to switch to os-checkfiles on Linux; rpm -V
--nomd5 disables the read checks. That's trivial enough to add to
the release right now.
* It still takes three minutes to do just a stat check (on
men-at-arms, which is kind of old hardware), which is probably too
long to run at reactivate. So maybe the decoupling proposal is a
good idea anyway.
3. Mozilla
We will put Mozilla 1.7 into the release, because it's the stable
branch, and because maybe we can use its Kerberos support to do neat
web things.
We would like to install an Xft-enabled Mozilla build, but Bob reports
that under Solaris, even with the third/redhat-fonts upgrade, it uses
an ugly serif font in the menus. Bob will look into that.
