Re: soak time (was: Ann: Subversion 0.35.1 released)

On Mon, 2003-12-22 at 21:56, solo turn wrote:
> --- kfogel@collab.net wrote:
> > I don't think the "soak time" system itself is broken, we just had
> > some human error this time. No system is going to eliminate that
> > completely, but the soak time is still valuable...

I agree.

> i like gregs proposition, as human errors are possible and do not matter.

In this case I think it certainly does matter.

> for deployment, we always keep a week soak time,

We?

> which means if releasing takes a week soak time and we do another
> week, we are at two weeks now ...

The Subversion project doesn't take 3rd party deployment soak time into
account. The soak time is there so that the critical features that are
destined for a release aren't committed on the day of the deadline and
released the same day. Since we do branch on the day of the milestone,
testing can occur for an entire week to flush out issues. Basically
nothing but critical bugfixes should be included on the branch.

It helps when the people involved with the various aspects of the
release process would test on the branch day (or close after).

Sander

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Dec 23 00:57:24 2003