[fab] Fw: [Bug 174307] RPM 4.4.6 is available

seth vidal skvidal at linux.duke.edu
Sun Jul 9 23:04:39 UTC 2006


On Sun, 2006-07-09 at 15:49 -0400, Michael Tiemann wrote:
> On Sun, 2006-07-09 at 15:30 -0400, seth vidal wrote:
> > 
> > I'd rather our package mgmt direction be a bit more organized than
> > reactionary responses to something that gets added one afternoon.
> 
> I think there are two separate problems.  The first, which may well be a
> fait accompli, is "which version of rpm should be in fc6?"  The second,
> which concerns me greatly, is "how/why the heck did we act/not act so
> that we'd have no choice in this matter for fc6?"
> 
> It looks to me from the time line of the bug report that we've had
> plenty of opportunity to hash out exactly how or whether we'd follow the
> upstream RPM path well in advance of the fc6 cutoff.  Heck, it could
> have made fc5!  But we, Red Hat, did not.  And we did not say why not.
> And now it looks like it's too late, which means that a strategy of
> inaction and non-response worked to achieve a tactical agenda that
> somebody, I don't know who, is pursuing.  Bully for them.  But we owe it
> to ourselves and the community, whether or not we can change our
> decision about the rpm version packaged for fc6, to explain fully and
> faithfully exactly why we've chosen to extend our divergence from
> upstream.  I don't think an implicit "deal with it" is going to cut it.
> 

In my opinion the reasons we can give for not implementing it are:

1. the policy decisions for automated action are not simple and not
hashed out, yet.

2. the primary focus for fc5 was making anaconda and yum play nicely
together. Adding some changes to rpm would have simply been too much
work

3. rpm is a major, major component of how the distro does what it does,
changes to rpm should be ultraconservative which don't seem to be the
case with all the new bits being added to the 4.4.X releases as it goes
along (YAML for no apparent reason, for example). If we change rpm we
have to check a bunch of components b/c of how deeply rpm is rooted into
things.

This is only my opinion  - but the board is still discussing this on the
board list. I'm pretty sure it's on the agenda for the next conference
call. I'll make sure we come to an answer by then - however it is really
an issue for the fedora core steering committee/cabal and it sounds like
to me they've already decided for this release as is evidenced in the
rawhide trees.

Does this make sense to you?

-sv







More information about the fedora-advisory-board mailing list