FreeBSD-7.1, BETA2 or PRERELEASE

[mdh]

--- On Fri, 10/24/08, Jeremy Chadwick <koitsu at FreeBSD.org> wrote:
> From: Jeremy Chadwick <koitsu at FreeBSD.org>
> Subject: Re: FreeBSD-7.1, BETA2 or PRERELEASE
> To: "Masoom Shaikh" <masoom.shaikh at gmail.com>
> Cc: freebsd-questions at freebsd.org
> Date: Friday, October 24, 2008, 6:26 AM
> On Fri, Oct 24, 2008 at 03:07:32PM +0530, Masoom Shaikh
> wrote:
> > Hi folks,
> > y'day I csuped the src and built installed the
> kernel from RELENG_7
> > I was expecting FreeBSD-BETA2 in output of `uname -a`
> > it is still -PRERELEASE, is it by decision or I have
> to change something ?
> >
> > I greped /usr/src for PRERELEASE but cud not locate
> it. I guess release
> > engineering team does that. comments ?
> 
> This question keeps coming up.
> 
> http://lists.freebsd.org/pipermail/freebsd-questions/2008-October/184992.html
> 
> RELENG_7 == PRERELEASE.  There is no "BETA2" tag
> to follow.
> 
> No one is sure at this point where the "BETA2"
> string has come from
> (meaning why it was idealised or why it's being used). 
> I'm of the
> belief that it's something Ken is hand-hacking in
> newvers.sh before
> building + making ISO releases and putting them up on the
> mirrors.
> And I am also of the opinion that this should stop, and we
> should simply
> name the releases PRERELEASE-YYYYMMDD to signify the build
> date.

It seems likely.  I've only ever seen -PRERELEASE and -STABLE, when tracking "RELENG_[0-9]" branch.  On the other hand, I have seen -RELEASE, -BETA, -RC, etc, when installing from media.  

Perhaps differentiating these isn't a bad idea, however, when it comes to uname output in PR's, despite the queries it generates over here.  A media install can always be safely assumed to be a given set of code, while if someone is tracking a branch via cvsup, the build time would show up in uname output, however the user may still need to be queried for rcsid's or asked to cvsup to the latest if the issue is considered to possibly be a base system and/or kernel code issue.  

It's probably worth discussion and consideration, though.  I don't know if/how useful the utility of the current naming conventions are to folks trying to solve potential code bug PRs.  

- mdh