Re: delay of the full etch freeze
Steve Langasek <email@example.com> writes:
> On Thu, Oct 12, 2006 at 10:22:43PM +0200, Petter Reinholdtsen wrote:
>> [Charles Plessy]
>> > The rationale is that the 8th is "old freeze deadline minus 10
>> > days", so it was not completely unreasonnable to take this day as
>> > the deadline for having new packages in Etch.
>> I find this completely unreasonable. If someone waited that late in
>> the release process before uploading a package they knew would have to
>> go through NEW, they can not expect the package to make it into Etch.
>> New packages should have had at least a few weeks in unstable to allow
>> problems to be detected before heading for testing.
>> So I would recommend against moving the freeze deadline to allow
>> packages in NEW to enter.
> Yes, this is my official position on the question (dunno about Andi's, I'm
> replying to email off-line at the moment and haven't checked with him, but I
> would guess his position is similar).
> The only packages in NEW that I'm inclined to worry about are those that fix
> release-critical bugs.
I think this is unrealistic, because we cannot predict NEW's
behavior. It doesn't follow that somebody "waited that late"; it may
well be instead that they did everything they could, and it was the
processing of NEW that waited a long time.
Sometimes packages require more than one trip through NEW, and the
release process cannot assume that people "waited that late" when it
was not they, but ftpmaster, that was waiting.