Is portsdb/portupgrade broken?

Matthew Seaman m.seaman at
Sun Nov 7 21:52:56 GMT 2004

Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sun, Nov 07, 2004 at 07:15:24PM +0000, Michael Abbott wrote:
> >>The bug is in the base system, not the ports.  ... The fix ... has been=
> >>applied to RELENG_4 ... So update to 4.10-STABLE ... and the problem wi=
> >>be fixed.
> >
> >Odd.  That's what I'm running (4.10p2).  I'll rebuild and try the other=
> >things you suggest.
> Ah.
> It seems that cvs tag RELENG_4 has the update, but RELENG_4_10 (which is=
> what I have selected in my cvsupfile) does not.  Ouch.
> Indeed, I just found your message on this very topic:
> So, I'd like to ask your advice here.  Really I want the stability of=20
> 4.10, but I want this bug fix; also, I don't want to patch my own sources=
> Was there a compelling reason for not incorporating the patch to=20
> bt_split.c into current releases, in particular into 4.10?

I'm not sure why the patch to bt_split.c wasn't considered of
sufficient importance that it should have been applied to
4.10-RELEASE.  That's up to the Release Engineering and Security teams
to decide.
> What unhappy consequences am I going to risk by changing my cvsup tag fro=
> RELENG_4_10 to RELENG_4?

It depends what you're doing with the system in question.  If it's an
important server that your livelyhood depends on having running 24x7,
then you need to think long and hard about the risks vs. the benefits
of switching to 4.10-STABLE.  If it's just a home system, then I'd say
make the jump to 4.10-STABLE.  You won't regret it.  I've been
tracking RELENG_3 originally and then RELENG_4 throughout most of
their existences, and I don't think I've /ever/ run into a serious
problem[1] because of it.  That's partly luck -- there have been one or
two occasions where an incautious commit made 4.x-STABLE uncompilable,
but those were fixed within hours.

Besides, development on the 4.x series is winding down.  At the moment
it's behaving de facto pretty close to what the -RELEASE branches are
meant to be like -- and its been particularly quiet over the last few
weeks, what with getting 5.3 out of the door.  There is a 4.11-RELEASE
planned, I believe, which will lead to a bit more activity on the
branch before that happens, and then not a lot more.

Basically, 4.10-STABLE is a pretty minimal risk usually, and it's even
less than that right now.



[1] Not counting the stuff I've managed to break through my very own

Dr Matthew J Seaman MA, D.Phil.                       26 The Paddocks
                                                      Savill Way
PGP:         Marlow
Tel: +44 1628 476614                                  Bucks., SL7 1TH UK

Content-Type: application/pgp-signature
Content-Disposition: inline

Version: GnuPG v1.2.6 (FreeBSD)



More information about the Ukfreebsd mailing list