Re: errata patches vs patch branch

[Dave Taira <bodhi@hagakure.org>]

On Fri, 21 Feb 2003, tim smith wrote:

> > It will give you all the latest updates that are
> >
> > critical
> > and/or trivial
> > and/or deemed worthwhile feeding out
>
> so i take it all patches which have a practical application get an errata
> patch?

No. If you look at errata.html, out of the 6 patches for 3.2, 4 are labeled
security fix, and two are labeled reliability fix. None are labeled
performance improvement, feature add, or anything else. Looking back...
3.0 errata includes two installation issues, 2.9 errata includes a
documentation fix and an installation fix. So those are the kinds of things
that get errata patches.

> so far i have been applying .patches to my home systems and it runs
> great, i can expect the same reliablity and ease of use from the patch
> branch?

Yes. As Theo says below, the developers aim to never break -stable.

> my big fear is updating to suddenly find something is broken.

I've been running -release and -stable (upgrading as appropriate) for
several years now, and have never had -stable break anything for me.

> > (we use an internal discussion to decide what goes out, and we are not
> > very ameniable towards people suggesting what should be included;
> > since doing this stable work takes a lot of developer effort; our goal
> > is to NEVER break a stable tree, and that means: full testing must
> > happen)

It's ultimately your call. But IMHO, using cvs to keep up with -stable
is easier (and more easily automated) than keeping up with individual
patches.

[ Dave Taira <bodhi@hagakure.org>                2003.02.21/00:30:32 PST ]
[ Morlock for Hire                                                       ]
[ All we're looking for is a fair value for exploitation of our          ]
[ members' rights.                                      --Chris Amenica, ]
[                         ASCAP Assistant VP of New Media and Technology ]