use disttag ".1" for devel to avoid confusion (was: Re: Plan for tomorrow's (20070604) Release Engineering meeting)

[Axel Thimm]

On Mon, Jun 04, 2007 at 09:27:31PM +0200, Michael Schwendt wrote:
> The picture of when exactly the mass-rebuild would happen and how
> much the the maintainers would be involved, is way too
> blurred. Talking about QA and lots of automated tests, we're not
> there yet.

Right, so let's not test at all then. Close our eyes and ship a
product that has build stamps from over a period of seven months all
over. So pray instead of test?

> How do a devel cycle and a test period fit into this? If we test
> previously built packages on the road to a final product

... like for example 7 months ago, e.g. on FC6 ...

> that we want to ship, why do we rebuild packages although nothing
> wrong has been found with the binaries?

Sure, let's ship the binaries and have the users find out.

> Why take the risk of replacing working binaries (which are assumed
> to work fine unless there are PRs) with theoretically we should
> re-test from scratch?

Anything is assumed to work fine until you get both (or more) pieces
shipped back. There is a risk? Risk management says: "break it within
a controled environment", e.g. break it while it is still labeled
rawhide, not when it is at the user (bananaware).

Furthermore this isn't Schrödinger's cat, where the bug is only there
if we look at it.

> The dist tag alone is no sufficient reason to mass-rebuild an entire
> distribution.

No, that's not the reason at all.
-- 
Axel.Thimm at ATrpms.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/fedora-maintainers/attachments/20070604/6aa2bab0/attachment.sig>