----- Original Message -----
-----BEGIN PGP SIGNED MESSAGE-----
El Sat, 9 Nov 2013 16:26:14 -0700
Kevin Fenzi <kevin(a)scrye.com> escribió:
> On Sun, 10 Nov 2013 00:00:42 +0100
> Christoph Wickert <christoph.wickert(a)gmail.com> wrote:
> > Dennis,
> > with all due respect, but I think your attitude it fundamentally
> > wrong.
> > You are the Fedora release manager. It's your job to communicate
> > deadlines, to reach out to the people affected and nag them. You are
> > getting paid to do this, but you expect volunteers to ask you? Seems
> > you are living in an upside-down world.
> Perhaps you are thinking of the Fedora Program manager here?
As Kevin said, the person responsible for bugging people and
organising the schedule is Jaroslav.
Yeah, that's me but I've never been asked to put it into schedule. And I'd
be more than happy to do so, even do that bugging part.
I'd say the Go/No-Go should be the break point to release/not to release as
for primary offering. It could be a part of the Go/No-Go meeting to state
release readiness of all deliverables we have (based on the sign offs or
directly in the meeting?).
Do we have anything as
for marking tests passed or just http://fedoraproject.org/wiki/Releases/20/Spins?
My Job is to make and deliver what is supposed to be made and
And thank you for that!
I'm not going to comment what's written bellow, I think Dennis covered it
pretty well. I'm sorry if I missed the spis process, I know it was happening
and now I know it's enforced - so let's try to find the way how to put it
into schedule, communicate it better. It's definitely worth having it, and I'd
be more than happy to help with it!
> > We had the same situation with F19 and someone (I don't
> > who. Matthew? Adam? Peter?) put it in a nutshell: "Dennis, you
> > screwed it up. Stop arguing." AFAIR you promised you would do
> > better, but I don't see any change.
> I'm not sure what this is refering to?
> Also, I am not Dennis (obviously), but I've been trying to drive some
> of this process, so feel free to yell at me and also I'll try and
> answer your concerns.
I believe that communications have been much better this time around. I
have been expressing concerns with the spins process for some time now,
Kevin has been good enough to do the majority of the work to drive
change here. There has been discussions here and on devel list about
changing the process, FESCo signed off on it, spins owners have not
followed it. I am not trying to be harsh or put anyone down, but you
really need to meet me in the middle somewhere.
> > Please be so kind as to tell me:
> > 1. Did you send out reminders to the mailing lists? I don't see
> > anything on this list.
> We could send reminders I guess. Just "please test your spin"?
> Shouldn't that all be implied?
> I did send a email to test/devel/this list:
> [Fedora-spins] IMPORTANT, please read: Spins QA signoff for milestones
> > 2. Did you send out reminders to the spin owners directly?
> Do you have a up to date list of them? I looked a few months ago and
> on some of them I couldn't figure it out at all. I was hoping perhaps
> strangely that there would be a list all of them would be on... like
> this one?
As with Kevin, I don't know who are the owners for most of the Spins.
> > 3. What test cases do you think needs to be run on spins you
> > want to remove? All are based on Desktop spins which are already
> > tested. The only thing worth testing is
> > Testcase_desktop_menus, but it is due for final, not beta. QA has no
> > other testcases or criteria the spins need to meet.
> General functionality. That it boots, that the menus and such work.
> Basically avoiding things like:
> - f19 security spin shipping with all broken menus.
> - fN design suite where network didn't work.
> - over indented sizes basically all the time.
> Brown paper bag issues you can find using the image for a few
> minutes. Is that too much to ask for? one person in the world testing
> just one TC or RC compose of the thing?
As Kevin said all we asked for was some basic testing to make sure that
things were not obviously broken. we have shipped too many spins in the
past broken. we even explicitly make all spins at every TC and RC
through the whole cycle now.
> > 4. Do we actually need to run test cases for non-desktop spins?
> > They are not listed in the wiki, neither in the template nor
> > on any of the test results pages.
> Some of the spins have actually started writing their own, which is
> great, but really I just wanted a "does this thing have any glaring
> errors that make it not work at all" test..
echo what Kevin has said here.
> > I do understand that as the release engineer, you don't want to ship
> > spins that don't work. I agree we need to get rid of unmaintained
> > and untested spins. But none of the spins you want to remove seem
> > unmaintained for me. Frankly speaking it looks to me like you are
> > trying to reduce your workload by all means necessary.
> I can't speak for dennis, but he has said many times this isn't about
> his work, he can produce all the spins all the time.
Removing the spins is more work than shipping them. I ant to make sure
we don't ship things that are broken because we never bothered to test.
I've said this a few times. I am happy to make everything that people
want, but i want to make sure it works, again meet in the middle.
> > > Security and Games
> > > should have been removed from Alpha since there is no sign they
> > > were tested then.
> > Then we would need to remove LXDE, too.
> No, there was someone who tested it and signed off on it.
LXDE was signed off in the wiki and followed the policy, Games and
Security have not been signed off in the wiki at all. Design Suite was
signed off for Alpha but not Beta.
> > > When we decide on a gold release that means we have tested
> > > it and are sure it is good. you need to test before then.
> > We are not gold yet, are we?
> Yes. At thursdays go/no-go meeting we are 'go' for f20beta. It's setup
> and syncing to mirrors now.
as kevin said.
> > I apologize I am very direct in this mail. With your attitude, you
> > are driving volunteers away from Fedora. I know some who are close
> > to resign and I cannot accept this.
I have really tried to work with everyone. I'm sorry you don't like my
attitude, quite frankly I don't really like yours right now, but I am
not going to let that get in the way of trying to deliver a tested
Fedora. We did discuss the changes, we even had FESCo sign off on them.
It really is not at all about me. I really don't think its unfair that
people be asked to test and fill in some info in the wiki.
> Well, I'm personally open to ideas. I'm a bit frustrated too, because
> I tried really hard this cycle to post to devel and test and here
> about the proposed requirements. I got some very few replies. I don't
> know how better to communicate the changes and get people to agree to
> Then we always get "Oh no, we didn't know, try it again next cycle".
> Well, I did. I wanted to do this in f19, but backed off when people
> said it wasn't communicated. I posted a number of times at the early
> part of the cycle, I got FESCo to ack the changes.
> How can we improve the process? How on earth can I get spin
> maintainers to agree to changes! How can we stop shipping things 0
> people test?
I think we need to at the least replace Christoph as the spins wrangler
because while I believe he completely intends the very best he is way
over committed and is not getting the job done. Does Spins fit into the
Workstation Working group? do we need a Spins working group. We do need
to make sure that changes for the better happen.
I do think I will give the 3 spins a reprive and add them back to beta,
at the least it means more work for me to undo their removal, as well
as work from websites to add them back to the web pages and quite
possibly work for others that I am forgetting right now. its not a
simple thing to do that only effects me.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
-----END PGP SIGNATURE-----
spins mailing list