On Thu, 2018-09-06 at 10:13 -0600, Chris Murphy wrote:
On Wed, Sep 5, 2018 at 6:59 PM, Adam Williamson
<adamwill(a)fedoraproject.org> wrote:
> On Wed, 2018-09-05 at 12:14 -0600, Chris Murphy wrote:
> > My opinion, since there are few facts to go on to overcome the burden
> > stated in a written process and schedule for some time, is -1 FE. If
> > it was important enough to get 3.30 on actual Beta installation media,
> > it needed to be done before freeze. Not depend on a freeze exception.
> > That is definitely not how things are supposed to work.
>
> As it happens, it is how things have been working, though. We've
> granted freeze exceptions for GNOME megaupdates for many of the last
> several releases (I can go back and get precise numbers if you like).
That's OK. Besides, there's some chance I have voted +1 FE for a GNOME
megaupdate, not least of which is because:
> Notably, I can't recall a single instance where they broke the world.
> This is not something I can say about a lot of packages, so I think the
> desktop team deserves some credit and trust for that.
For sure. But that is orthogonal to freeze exception. It's like using
FE as some kind of Good Job sticker.
I wouldn't say it's orthogonal, at all. It's a significant factor. It's
true to say that we shouldn't grant something FE status *solely
because* we don't think it's likely to break anything terribly. It's
not *sufficient* justification in itself. But if there is, let's say, a
plausible case for "there are some good reasons to let this in Beta"
but *also* a plausible case for "...but it's kinda late and we don't
absolutely *need* it and it's change!", then the track record of that
team and that kind of change absolutely does factor into whether I'm
likely to vote +1.
And we all know there is increased probability that blocker bugs are
not found or are found later than otherwise, because of diverted
attention toward testing the megaupdate. There is no way of predicting
or assessing this, but logically it's true. And so this particular
usage of freeze exception ends up having more in common with craps,
than a well deserved Good Job sticker.
Eh. I don't think more people testing Workstation, for whatever reason,
is really going to be a *bad* thing. Maybe the increased attention
causes us to find a blocker we'd otherwise have missed?
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
http://www.happyassassin.net