EOL planning (and observations)

[David Eisenstein]

On Wed, 11 Jan 2006, Pekka Savola wrote:

> On Wed, 4 Jan 2006, Axel Thimm wrote:
> > according to original planning, the following EOL dates apply:
> >
> > RH7.3,RH9:   as long as community wants
> > FC1:	     until the release of FC4
> > FC2:	     until the release of FC5
> >
> > is this still reflecting the current setup? Is there some kind of
> > statistics on the master server about which distribution is still in
> > use?
> 
> So, what's our approach here?  Do we drop support for FC1 in all the 
> subsequent updates (those where packages haven't been proposed yet)? 
> Do we drop FC2 at the same time as we pick up FC3?

I am not sure Axel's reckoning of EOL dates for FC1 and FC2 is quite
right.  My understanding was that FC1 was to have been dropped when FC5
became reality (or nearly reality) and that we would pick up FC3 at that
time.  That way we would be supporting two Fedora Core EOL'ed distros at
any given time, as well as the two Red Hat distros.

However, Jesse Keating indicated, and I tend to agree (admittedly for
selfish reasons), that he is not willing to drop support for FC1 when FC5
Test 2 comes out and we're to inherit FC3:

(from <http://www.redhat.com/archives/fedora-legacy-list/2005-December/msg00018.html>, 
"Re: 8 more days 'til we inherit FC3; are we ready??; FWD:  Fedora Core 3
Status Update"):

On Thu, 15 Dec 2005 at 18:41:33 -0800, Jesse Keating wrote:
> On Thu, 2005-12-15 at 20:30 -0600, David Eisenstein wrote:
> > Hi,
> > 
> > In just a little over a week, we are scheduled to inherit Fedora Core 3.
> > I suppose at the same time, we will be dropping Fedora Core 1 (though I
> > wish we weren't -- we seem to have a lot more postings of parties 
> > interested in FC1 than we have interested in FC2, and we get more votes
> > and QA testing on FC1 than FC2 in bugzilla).   
> > 
> > Are we ready?  What do we need to prepare for maintaining FC3?
> 
> A few last minute syncups and maybe touching a config file or two and
> we'll be ready.
> 
> At this time I am not prepared to drop FC1.  We have a significant user
> base that is still supporting us in our FC1 tasks and I will not abandon
> them.

My observation is that we've had more people voting and participating and
expressing interest in FC1 issues than FC2 issues.  I am one continuing to
run FC1 and contribute.  So I am wondering -- if we need to drop a distro
at FC5 Test 2 time, should we drop FC2, for lack of interest (and lack of
*contributors*)?  Or do we have some kind of "silent majority" really
needing FC2 support to continue?

Another observation is that supporting even FOUR distros, with the amount
of work we have lately had going into the maintenance of these distros,
might be biting off more than we can chew ... or at least to chew and chew
well with current levels of package-building and QA participation.

When broad support for any given vulnerability is expressed here in our
forum, things get hopping:  Like what happened for PHP back in November.  
But in general people are not submitting packages for testing/QA very much
lately (I'm guilty of this too), and when packages *are* being submitted
for Quality Assurance, we seem to only have three or four people bothering
to vote in Bugzilla so we can push issues to the next stage and close
them.  Perhaps we're getting lazy??  **

"If it's to be, it's up to me."  What do you say, folks?

	Regards,

	David Eisenstein


------- footnote:

**  For example, we've had a Perl bug open across two Bugzillas and well
    over a year now.  (Admittedly, this is the first time Fedora Legacy
    has attempted to update Perl, and there have been some thorny issues
    to overcome.)  Perl is a pretty important package to many, because
    it's a work-horse language that most Linux distros cannot do without.
    But we've simply not seen people willing to go in and test and vote
    on these Perl packages.
    
    Currently it's in Red Hat Bugzilla, Bug #152845, and is in the
    "updates-testing" stage needing us to download and test the binaries.
    Pekka has stepped in to do QA on the RHL 9 and RHL 7.3 versions, but
    we still need testing on FC1 and FC2 versions.  I'd like to see these
    Perl packages published soon, because yet another vulnerability has
    cropped up and needs new packages (this is Bugzilla Bug #176731 --
    for CVE-2005-3962 format string vulnerability).

    (I would vote, but I am not supposed to:  I built the updates-testing
    packages, and it needs new eyes to see things I may have overlooked.)