Habeas check-ins

[Daniel Quinlan <qu...@pathname.com>]

I am -0.9 on these two changes going into 3.0.3 at this late date.  They
are not bug fixes and I do NOT want to release a 3.0.4 for any reason
other than a security update.  Do we have to do this here?

Daniel

-- 
Daniel Quinlan
http://www.pathname.com/~quinlan/

Re: Habeas check-ins

[Warren Togami <wt...@redhat.com>]

Michael Parker wrote:
> 
> I also have some free time for the next week or so, so I decided to
> take on the task, make a few people happy.  I certainly don't want to
> be troubleshooting AWL memory problems 2 years from now from someone
> using SA in a "stable" version of an OS.  Witness 2.20 in Debian
> stable, 2.55 in RHEL3 or 2.64 in SUSE Enterprise.  Hopefully by
> getting 3.0.3 out there and available it will make life easier down
> the road.
> 
> I really don't want to put in all this effort to find out a few weeks
> from now that an enhancement is causing problems and we'll have to
> live with those problems for a couple of years to come.
> 
> 
>>Also in general, discussions/votes ought to go in the ticket, otherwise
>>it's too hard to track where the conversations are taking place.
>>
> 
> 
> Yeah, I apologize, I should have made my veto more public.  I'll be
> sure to do so much earlier in the process next time.

According to quinlan it isn't tested well and wouldn't change scores 
much, so I vote to not take the risk.  Pure bug fixes only please.

Warren Togami
wtogami@redhat.com

Re: Habeas check-ins

[Michael Parker <pa...@pobox.com>]

On Tue, Apr 26, 2005 at 06:24:01PM -0400, Theo Van Dinter wrote:
> On Tue, Apr 26, 2005 at 04:41:39PM -0500, Michael Parker wrote:
> > > I am -0.9 on these two changes going into 3.0.3 at this late date.  They
> > > are not bug fixes and I do NOT want to release a 3.0.4 for any reason
> > > other than a security update.  Do we have to do this here?
> > 
> > Agreed, the primary motivation for this release is a rock solid final
> > release for 3.0.  Several distributions have been asking for this so
> > they can roll the latest into their releases.  Adding features is not
> > something that should be happening at this point in time.
> 
> The patches were supposed to be in 3.0.2, but got skipped for several
> reasons, all of them procedural.  I'm also not sure what "late date"
> means here -- 3998 has been in the 3.0.3 queue for as long as there's
> been a 3.0.3 queue.  It's not as if it's a last minute set of code or
> something that just got put in the queue last night.  There were no
> current vetos in the tickets as of this morning for either patch, FWIW.

Perhaps a reluctance to put new functionality into a maintenance
release.  Age of the code isn't the issue, it's the fact that it's an
enhancement to a stable code release.

Admittedly I should have made my veto more public and added it to the
bug.  It's there now.  I'm quite adamant about making the 3.0.3
release very solid.  It fixes several annoyances in 3.0.2 and I
believe it should provide some longevity for the 3.0 code line.

> Frankly, the whole 3.0.3 release is strange imo.  We all were saying,
> for several months now, that there was not going to be a 3.0.3 release
> and were focusing on 3.1 w/ an upgrade being the official "bug fix/etc"
> path.  That's why most things in the 3.0.3 queue have just sat there
> and haven't been commented on.  Now suddenly (for a week or two anyway)
> there's a strong desire to get 3.0.3 out?  Alright, enough venting.

Yes, I believe the intention was to focus our efforts on 3.1.  It just
so happens that recently several folks have been asking about a 3.0.3
and getting a stable release for some long lived distributions.

I also have some free time for the next week or so, so I decided to
take on the task, make a few people happy.  I certainly don't want to
be troubleshooting AWL memory problems 2 years from now from someone
using SA in a "stable" version of an OS.  Witness 2.20 in Debian
stable, 2.55 in RHEL3 or 2.64 in SUSE Enterprise.  Hopefully by
getting 3.0.3 out there and available it will make life easier down
the road.

I really don't want to put in all this effort to find out a few weeks
from now that an enhancement is causing problems and we'll have to
live with those problems for a couple of years to come.

> 
> Also in general, discussions/votes ought to go in the ticket, otherwise
> it's too hard to track where the conversations are taking place.
> 

Yeah, I apologize, I should have made my veto more public.  I'll be
sure to do so much earlier in the process next time.

Michael

Re: Habeas check-ins

[Theo Van Dinter <fe...@kluge.net>]

On Tue, Apr 26, 2005 at 04:41:39PM -0500, Michael Parker wrote:
> > I am -0.9 on these two changes going into 3.0.3 at this late date.  They
> > are not bug fixes and I do NOT want to release a 3.0.4 for any reason
> > other than a security update.  Do we have to do this here?
> 
> Agreed, the primary motivation for this release is a rock solid final
> release for 3.0.  Several distributions have been asking for this so
> they can roll the latest into their releases.  Adding features is not
> something that should be happening at this point in time.

The patches were supposed to be in 3.0.2, but got skipped for several
reasons, all of them procedural.  I'm also not sure what "late date"
means here -- 3998 has been in the 3.0.3 queue for as long as there's
been a 3.0.3 queue.  It's not as if it's a last minute set of code or
something that just got put in the queue last night.  There were no
current vetos in the tickets as of this morning for either patch, FWIW.

Frankly, the whole 3.0.3 release is strange imo.  We all were saying,
for several months now, that there was not going to be a 3.0.3 release
and were focusing on 3.1 w/ an upgrade being the official "bug fix/etc"
path.  That's why most things in the 3.0.3 queue have just sat there
and haven't been commented on.  Now suddenly (for a week or two anyway)
there's a strong desire to get 3.0.3 out?  Alright, enough venting.

As for the 3998 commit, I miscounted the vote that did occur, so that's
reverted now.  Sorry.

Also in general, discussions/votes ought to go in the ticket, otherwise
it's too hard to track where the conversations are taking place.

-- 
Randomly Generated Tagline:
"I don't even have to get dressed up for Halloween.  I go as me." - Judge Judy

Re: Habeas check-ins

[Michael Parker <pa...@pobox.com>]

On Tue, Apr 26, 2005 at 02:36:09PM -0700, Daniel Quinlan wrote:
> I am -0.9 on these two changes going into 3.0.3 at this late date.  They
> are not bug fixes and I do NOT want to release a 3.0.4 for any reason
> other than a security update.  Do we have to do this here?

Agreed, the primary motivation for this release is a rock solid final
release for 3.0.  Several distributions have been asking for this so
they can roll the latest into their releases.  Adding features is not
something that should be happening at this point in time.

Michael