Support for 5.x (Was: Re: What about BIND 9.3.4 in FreeBSD in base system ?)

[Doug Barton]

Chris Marlatt wrote:

> Again, why would you expect someone to have already upgraded when they 
> have more than a year of advertised support left on a production release?

I think that there is a misunderstanding here about what "support" 
means. This discussion has been had in detail on several other lists 
of late, so let me give you the highlights. There are three main issues.

1. Support for security issues. The secteam has pledged to support the 
5.5-RELEASE through 31 May 2008.
2. Ports. The ports team generally pledges to support a release as 
long as the security team does, but the idea of dropping support early 
for RELENG_5 has been discussed.
3. New development, performance, features, etc. It is a virtual 
certainty that none of this will happen in RELENG_5. It is certain (by 
definition) that no new features will be backported that require ABI 
breakage, which severely limits what can come back into this branch 
already.

> I personally have very few 5.x systems left, primarily because I've been 
> trying to heed the warnings, but seeing how 5 series is being fast 
> tracked into retirement makes me extremely suspicious of what is to 
> happen to 6 series when 7 is released and considered production.

I can't make any iron clad assurances here, but I can say that this is 
unlikely to happen, because of why a lot of us want to drop support 
for 5.x. Namely that it has a lot of issues that cannot be fixed 
without breaking the ABI. In many key areas, 6.x is light years ahead, 
and is going to stay that way. There are some incremental improvements 
in 7-current right now that will make it attractive to some users, but 
if you're looking for something to install and keep supported for a 
longer time period, 6.x is going to be it.

> I'm 
> sure many other people wonder the same thing and look at the lengthy 
> support for 4 series which lasted 7,... 8 years and have come to expect 
> something similar for future releases.

Which is why we're working so hard to disabuse people of that notion.

> Whereas I'm certainly not going 
> to say progress is evil I will admit that the FreeBSD I see today is not 
> the same one from yesteryear.

Nor will it be the same tomorrow. Such is life in a volunteer project.

> Now, I can clearly understand and appreciate the burden that, as of 
> yesterday, 3 active versions

Actually it was 4 active versions, including 7-current.

> can impose on the development team but why 
> pass part of that burden onto a user base that's done nothing but 
> embraced the products produced by its efforts?

This is where that whole "volunteer project" thing comes in again. 
With a finite set of resources, we have to be realistic in terms of 
how thin we can spread them. Right now the rough order of priority is 
6-stable, 7-current, 5-stable. It is of course up to you to decide how 
to manage your own resources, but please don't expect magical things 
to happen in 5-stable just because you'd like them to. :)

Doug

-- 

     This .signature sanitized for your protection