Re: Debian 3.0r1
On Thu, 25 Jul 2002, Martin Schulze wrote:
> Make the bugs a security problem or move Easter and Christmas to the
> same date.
Allowing for the precession of the equinoxes that should happen in (very
approximately) 7,000 years. So you're saying you'll consider it for
> More serious:
> Come on people. What makes you think that I'm going be more sloopy
> with the Woody release, that took us more time than any other release,
> and change the policy on revisions of the stable Debian release?
> Once a Debian distribution is released as stable, the release is meant
> to be stable, except for security and very serious updates. The
> outcome is that merely security updates and updates to very serious
> (please don't confuse this with the BTS severity ``serious'') problems
> will make its way into the stable release through a
> point-release/revision. Such problems refer to potential data loss,
> uninstallable package or unusable package.
Unlike Russell, the updates I had in mind mainly involve packaging errors
on my part not new code.
> The whole point of having well tested packages, a well tested and
> integrated stable distribution, and keeping the stable release as
> stable as possible and without too much changes would be moot if
> random packages were able to get updated and added.
I agree with this on thw whole but I feel sometimes we take this to absurd
lengths. If we keep things which are clearly broken how is that more
stable than newer code which only potentially might break things?
And I doubt Russell was suggesting adding random packages. There should
be some mechanism by which we as a group can decide a slightly more
substantial update can go into stable if deemed necessary. "No new code"
is a nice and simple formulation but simple doesn't necessarily mean
Jaldhar H. Vyas <email@example.com>
It's a girl! See the pictures - http://www.braincells.com/shailaja/
To UNSUBSCRIBE, email to firstname.lastname@example.org
with a subject of "unsubscribe". Trouble? Contact email@example.com