On Wed, 2002-08-21 at 13:50, Marc G. Fournier wrote:
> On Wed, 21 Aug 2002, Bruce Momjian wrote:
>
> >
> > We learned a few lessons from previous releases. First, don't delay
> > the beta by days/weeks that drag on. Delay one month at a time.
> > Second, don't decide on a further delay the day before you are going to
> > go beta. Multiple short-period delays and delays that happen at the
> > last minute cause too many stops/starts for developers to be effective,
> > so...
> >
> > If we are going to delay beta, we should decide now, not at the end of
> > August, and the delay should be until the end of September. The big
> > question is whether we have enough material to warrant a delay.
>
> Beta goes down in 1 week ... if we follow what we had talked about before,
> within a short period of time after beta, we should be able to let ppl
> dive into working on v7.4 (or 8.0, whatever we decide to call it) ... but
> let's try and stick to a timeline for once, else we are going to hit the
> same as the last *very* extended release ...
Agreed. If patches are applied to the 7.4 branch as fast as normal,
then maybe 7.4 will only be 6 months out with well tested Windows, PIT,
etc. code that gets applied this October.
Whats the intended branchpoint? Beta with less than 5 patches? 3rd
beta start period? Less than 100 lines changed between betas?
Where is the reasonable point where double patching isn't as annoying as
waiting to apply new work?